> **Building with AI coding agents?** Install the authstack plugin with one command. This equips your agent with accurate Scalekit implementation patterns.
>
> **Recommended** (the `-y` flags skip prompts, so the command runs without an interactive terminal):
> ```bash
> npx -y @scalekit-inc/cli setup -y
> ```
>
> Global:
> ```bash
> npm install -g @scalekit-inc/cli
> scalekit setup -y
> ```
>
> Supports Claude Code, Cursor, GitHub Copilot, Codex + skills for other Agent Skills-compatible agents in `~/.agents/skills`.
> Skills: integrate-agentkit, implement-saaskit, add-mcp-oauth, implement-sso, implement-scim.
> [Full setup guide](https://docs.scalekit.com/dev-kit/build-with-ai/)

---

# Connect AI agents to Zep

The Zep connector lets your AI agent act in each user's Zep account. Each user connects their own Zep API key once, and Scalekit sends it with every call, so your agent never handles credentials. It comes with 60 tools.

**Authentication:** API Key
**Categories:** AI, Databases
**Tools:** 60: 38 read, 16 write, 6 destructive
**Users sign in with:** API key
**Built by:** Scalekit connector
**Try it:** [Playground in the Scalekit dashboard](https://app.scalekit.com/ws/signup?sk_intent=playground&provider=ZEP)

## Setup

1. ### Install the SDK

   **Node.js**

   ```bash
   npm install @scalekit-sdk/node dotenv
   ```

   **Python**

   ```bash
   pip install scalekit-sdk-python python-dotenv
   ```

2. ### Set your credentials

   Add your Scalekit credentials to your `.env` file. Find values in **[app.scalekit.com](https://app.scalekit.com)** > **Developers** > **API Credentials**.

   ```sh title=".env"
   SCALEKIT_ENVIRONMENT_URL=<your-environment-url>
   SCALEKIT_CLIENT_ID=<your-client-id>
   SCALEKIT_CLIENT_SECRET=<your-client-secret>
   ```

3. ### Create the Zep connection

   In **AgentKit > Connections**, create a Zep connection. The name you give it is the `connection_name` your code passes. See [Configure connections](/agentkit/connections/).

4. ### Authorize a user and make your first call

   **Node.js** (`quickstart.mts`)

   ```typescript
   import { ScalekitClient } from '@scalekit-sdk/node'
   import 'dotenv/config'
   import { createInterface } from 'node:readline/promises'

   const scalekit = new ScalekitClient(
     process.env.SCALEKIT_ENVIRONMENT_URL,
     process.env.SCALEKIT_CLIENT_ID,
     process.env.SCALEKIT_CLIENT_SECRET,
   )
   const actions = scalekit.actions

   const connector = 'zep'
   const identifier = 'user_123'

   // Generate an authorization link for the user
   const { link } = await actions.getAuthorizationLink({ connectionName: connector, identifier })
   console.log('Authorize Zep:', link)
   const rl = createInterface({ input: process.stdin, output: process.stdout })
   await rl.question('Press Enter after authorizing...')
   rl.close()

   // Make your first call
   const result = await actions.executeTool({
     connector,
     identifier,
     toolName: 'zep_graph_lookup',
     toolInput: {},
   })
   console.log(result)
   ```

   ```bash
   npx tsx quickstart.mts
   ```

   **Python** (`quickstart.py`)

   ```python
   import os
   from scalekit import ScalekitClient
   from dotenv import load_dotenv
   load_dotenv()

   scalekit_client = ScalekitClient(
       env_url=os.getenv("SCALEKIT_ENVIRONMENT_URL"),
       client_id=os.getenv("SCALEKIT_CLIENT_ID"),
       client_secret=os.getenv("SCALEKIT_CLIENT_SECRET"),
   )
   actions = scalekit_client.actions

   connection_name = "zep"
   identifier = "user_123"

   # Generate an authorization link for the user
   link_response = actions.get_authorization_link(
       connection_name=connection_name,
       identifier=identifier,
   )
   print("Authorize Zep:", link_response.link)
   input("Press Enter after authorizing...")

   # Make your first call
   result = actions.execute_tool(
       tool_input={},
       tool_name="zep_graph_lookup",
       connection_name=connection_name,
       identifier=identifier,
   )
   print(result)
   ```

   ```bash
   python quickstart.py
   ```

   Each user opens the link once and enters their Zep credentials there. See [Authorize a user](/agentkit/tools/authorize/) for the full flow and statuses.

## Tools

Pass the exact name to `execute_tool`, with an input like each tool's example. To ask for a tool that's missing, use the [request form](https://scalekitsupport.portal.usepylon.com/forms/request-a-connector-tool).

**Node.js**

```typescript
const result = await actions.executeTool({
  toolName: 'zep_edge_get',
  toolInput: {
    edge_uuid: '<edge_uuid>',
    graph_uuid: '<graph_uuid>',
  },
  connector: 'zep',
  identifier: 'user_123',
})
```

**Python**

```python
result = actions.execute_tool(
    tool_name="zep_edge_get",
    tool_input={
        "edge_uuid": "<edge_uuid>",
        "graph_uuid": "<graph_uuid>",
    },
    connection_name="zep",
    identifier="user_123",
)
```

`result.data` is the app's response as JSON, and `result.execution_id` (`executionId` in Node.js) is the ID of the call.

### `zep_edge_get`

Get Edge · Read-only

Get one edge with its fact, source and target nodes and validity times. Returns the edge object. Get UUIDs from zep_edges_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `edge_uuid` (`string`, required): UUID of the edge. Get it from zep_edges_list or zep_graph_edges_search. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"edge_uuid":"<edge_uuid>","graph_uuid":"<graph_uuid>"}`

### `zep_edges_list`

List Edges · Read-only

List a graph's edges (facts), a page at a time, optionally filtered. Returns edge objects with their fact, nodes and validity times. Use zep_graph_edges_search for a text query. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `order` (`string`, optional): Sort direction: asc or desc (default desc). Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort key: uuid (default), created_at, or valid_at. Example: created_at. One of: `uuid`, `created_at`, `valid_at`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_episode_get`

Get Episode · Read-only

Get one episode with its content and processing status. Returns the episode object. Get UUIDs from zep_episodes_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `episode_uuid` (`string`, required): UUID of the episode. Get it from zep_episodes_list. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"episode_uuid":"<episode_uuid>","graph_uuid":"<graph_uuid>"}`

### `zep_episodes_list`

List Episodes · Read-only

List a graph's episodes, newest first, a page at a time. Returns episode objects with content and processing status. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `order` (`string`, optional): Sort direction: asc or desc. Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort field. Example: created_at. One of: `created_at`, `uuid`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_graph_context_get`

Assemble Graph Context · Read-only

Search a graph with a query and assemble the results into a context block for an LLM prompt. Returns the context string. Use the zep_graph_*_search tools for raw results and zep_thread_context_get for a thread. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `query` (`string`, required): Natural-language search query, for example What plan is Jane on? Example: What plan is Jane on?.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `include_results` (`boolean`, optional): When true, includes the raw graph results selected for the context block. Example: true.
- `max_characters` (`integer`, optional): The maximum number of characters in the assembled context block. Example: 10.
- `recency_bias` (`string`, optional): Adjusts result selection to favor more recent graph data. Example: off. One of: `off`, `mild`, `strong`.
- `template_uuid` (`string`, optional): The UUID of a context template used to render the context block. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.

Example input: `{"graph_uuid":"<graph_uuid>","query":"<query>"}`

### `zep_graph_edges_search`

Search Graph Edges · Read-only

Search a graph's facts (edges between entities) by a text query, with optional filters and reranking. Returns the best-matching edges with scores. Use zep_graph_context_get for a ready-made prompt block. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `query` (`string`, required): Natural-language search query, for example What plan is Jane on? Example: What plan is Jane on?.
- `bfs_origin_node_uuids` (`array`, optional): Nodes used as BFS origins for graph-distance-aware retrieval. Example: ["8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e"]
- `center_node_uuid` (`string`, optional): The node to rank results by distance from. Required when reranker is node_distance. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filter_low_quality_results` (`boolean`, optional): The default is false. When true, Zep removes low-quality results at the knee of the cross-encoder score curve. The filter never removes a result with a score of 0.9 or more. It keeps at least 5 results, or all results when fewer than 5 are available. The filter only removes results, so a page still returns up to page_size results. Use this field only with reranker cross_encoder. It is not supported for episode search. Invalid combinations return a 400 error. The cutoff is approximate, not guaranteed. When the cross-encoder does not score the results, for example after a timeout, Zep... Example: true.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `mmr_lambda` (`number`, optional): The diversity weighting used for maximal marginal relevance reranking. Required when reranker is mmr. Example: 0.5.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.
- `reranker` (`string`, optional): How results are reranked: rrf (default), mmr, node_distance, episode_mentions or cross_encoder. Example: rrf. One of: `rrf`, `mmr`, `node_distance`, `episode_mentions`, `cross_encoder`.

Example input: `{"graph_uuid":"<graph_uuid>","query":"<query>"}`

### `zep_graph_episodes_search`

Search Graph Episodes · Read-only

Search a graph's episodes (the raw data added to the graph) by a text query, with optional filters and reranking. Returns the best-matching episodes with scores. Use zep_graph_context_get for a ready-made prompt block. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `query` (`string`, required): Natural-language search query, for example What plan is Jane on? Example: What plan is Jane on?.
- `bfs_origin_node_uuids` (`array`, optional): Nodes used as BFS origins for graph-distance-aware retrieval. Example: ["8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e"]
- `center_node_uuid` (`string`, optional): The node to rank results by distance from. Required when reranker is node_distance. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filter_low_quality_results` (`boolean`, optional): The default is false. When true, Zep removes low-quality results at the knee of the cross-encoder score curve. The filter never removes a result with a score of 0.9 or more. It keeps at least 5 results, or all results when fewer than 5 are available. The filter only removes results, so a page still returns up to page_size results. Use this field only with reranker cross_encoder. It is not supported for episode search. Invalid combinations return a 400 error. The cutoff is approximate, not guaranteed. When the cross-encoder does not score the results, for example after a timeout, Zep... Example: true.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `mmr_lambda` (`number`, optional): The diversity weighting used for maximal marginal relevance reranking. Required when reranker is mmr. Example: 0.5.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.
- `reranker` (`string`, optional): How results are reranked: rrf (default), mmr, node_distance, episode_mentions or cross_encoder. Example: rrf. One of: `rrf`, `mmr`, `node_distance`, `episode_mentions`, `cross_encoder`.

Example input: `{"graph_uuid":"<graph_uuid>","query":"<query>"}`

### `zep_graph_get`

Get Graph · Read-only

Get one graph by UUID. Returns the graph with its name, description and owner. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_graph_instructions_get`

Get Graph Instructions · Read-only

Get the custom extraction instructions set on a graph. Returns the instructions. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_graph_lookup`

Look Up Graph by Developer ID · Read-only

Find a graph by the developer ID (your own identifier) you gave it. Returns the matching graph. Use zep_graph_get with a Zep UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_id` (`string`, optional): The developer-assigned graph ID to resolve to a UUID. Mutually exclusive with user_id and thread_id. Example: product-docs.
- `thread_id` (`string`, optional): The developer-assigned thread ID to resolve to a UUID. Mutually exclusive with user_id and graph_id. Example: support-chat-2026-01-15.
- `user_id` (`string`, optional): The developer-assigned user ID to resolve to a UUID. Mutually exclusive with thread_id and graph_id. Example: auth0|64f1c2.

Example input: `{}`

### `zep_graph_nodes_search`

Search Graph Nodes · Read-only

Search a graph's entities (nodes) by a text query, with optional filters and reranking. Returns the best-matching nodes with scores. Use zep_graph_context_get for a ready-made prompt block. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `query` (`string`, required): Natural-language search query, for example What plan is Jane on? Example: What plan is Jane on?.
- `bfs_origin_node_uuids` (`array`, optional): Nodes used as BFS origins for graph-distance-aware retrieval. Example: ["8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e"]
- `center_node_uuid` (`string`, optional): The node to rank results by distance from. Required when reranker is node_distance. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filter_low_quality_results` (`boolean`, optional): The default is false. When true, Zep removes low-quality results at the knee of the cross-encoder score curve. The filter never removes a result with a score of 0.9 or more. It keeps at least 5 results, or all results when fewer than 5 are available. The filter only removes results, so a page still returns up to page_size results. Use this field only with reranker cross_encoder. It is not supported for episode search. Invalid combinations return a 400 error. The cutoff is approximate, not guaranteed. When the cross-encoder does not score the results, for example after a timeout, Zep... Example: true.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `mmr_lambda` (`number`, optional): The diversity weighting used for maximal marginal relevance reranking. Required when reranker is mmr. Example: 0.5.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.
- `reranker` (`string`, optional): How results are reranked: rrf (default), mmr, node_distance, episode_mentions or cross_encoder. Example: rrf. One of: `rrf`, `mmr`, `node_distance`, `episode_mentions`, `cross_encoder`.

Example input: `{"graph_uuid":"<graph_uuid>","query":"<query>"}`

### `zep_graph_observations_search`

Search Graph Observations · Read-only

Search a graph's observations (insights Zep derived from the data) by a text query, with optional filters and reranking. Returns the best-matching observations with scores. Use zep_graph_context_get for a ready-made prompt block. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `query` (`string`, required): Natural-language search query, for example What plan is Jane on? Example: What plan is Jane on?.
- `bfs_origin_node_uuids` (`array`, optional): Nodes used as BFS origins for graph-distance-aware retrieval. Example: ["8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e"]
- `center_node_uuid` (`string`, optional): The node to rank results by distance from. Required when reranker is node_distance. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filter_low_quality_results` (`boolean`, optional): The default is false. When true, Zep removes low-quality results at the knee of the cross-encoder score curve. The filter never removes a result with a score of 0.9 or more. It keeps at least 5 results, or all results when fewer than 5 are available. The filter only removes results, so a page still returns up to page_size results. Use this field only with reranker cross_encoder. It is not supported for episode search. Invalid combinations return a 400 error. The cutoff is approximate, not guaranteed. When the cross-encoder does not score the results, for example after a timeout, Zep... Example: true.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `mmr_lambda` (`number`, optional): The diversity weighting used for maximal marginal relevance reranking. Required when reranker is mmr. Example: 0.5.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.
- `reranker` (`string`, optional): How results are reranked: rrf (default), mmr, node_distance, episode_mentions or cross_encoder. Example: rrf. One of: `rrf`, `mmr`, `node_distance`, `episode_mentions`, `cross_encoder`.

Example input: `{"graph_uuid":"<graph_uuid>","query":"<query>"}`

### `zep_graph_ontology_get`

Get Graph Ontology · Read-only

Get the custom entity and edge types (ontology) set on a graph. Returns the ontology. Use zep_project_ontology_get for the project default. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_graph_thread_summaries_search`

Search Graph Thread Summaries · Read-only

Search a graph's thread summaries by a text query, with optional filters and reranking. Returns the best-matching thread summaries with scores. Use zep_graph_context_get for a ready-made prompt block. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `query` (`string`, required): Natural-language search query, for example What plan is Jane on? Example: What plan is Jane on?.
- `bfs_origin_node_uuids` (`array`, optional): Nodes used as BFS origins for graph-distance-aware retrieval. Example: ["8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e"]
- `center_node_uuid` (`string`, optional): The node to rank results by distance from. Required when reranker is node_distance. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filter_low_quality_results` (`boolean`, optional): The default is false. When true, Zep removes low-quality results at the knee of the cross-encoder score curve. The filter never removes a result with a score of 0.9 or more. It keeps at least 5 results, or all results when fewer than 5 are available. The filter only removes results, so a page still returns up to page_size results. Use this field only with reranker cross_encoder. It is not supported for episode search. Invalid combinations return a 400 error. The cutoff is approximate, not guaranteed. When the cross-encoder does not score the results, for example after a timeout, Zep... Example: true.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `mmr_lambda` (`number`, optional): The diversity weighting used for maximal marginal relevance reranking. Required when reranker is mmr. Example: 0.5.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.
- `reranker` (`string`, optional): How results are reranked: rrf (default), mmr, node_distance, episode_mentions or cross_encoder. Example: rrf. One of: `rrf`, `mmr`, `node_distance`, `episode_mentions`, `cross_encoder`.

Example input: `{"graph_uuid":"<graph_uuid>","query":"<query>"}`

### `zep_graphs_list`

List Graphs · Read-only

List the standalone graphs in the project, a page at a time. Returns graph objects with their UUIDs. Use zep_graph_lookup to find a graph by your own ID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `order` (`string`, optional): asc or desc. Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort field. Example: created_at. One of: `created_at`, `name`, `uuid`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.
- `search` (`string`, optional): Filters results to graphs whose name, description, or graph ID contains this text. Example: quarterly report.

Example input: `{}`

### `zep_node_get`

Get Node · Read-only

Get one node with its name, labels, summary and attributes. Returns the node object. Get UUIDs from zep_nodes_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `node_uuid` (`string`, required): UUID of the node. Get it from zep_nodes_list or zep_graph_nodes_search. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.

Example input: `{"graph_uuid":"<graph_uuid>","node_uuid":"<node_uuid>"}`

### `zep_node_neighbors_list`

List Node Neighbors · Read-only

List the nodes connected to a node, with the edges between them. Returns neighbor nodes and edges. Get UUIDs from zep_nodes_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `node_uuid` (`string`, required): UUID of the node. Get it from zep_nodes_list or zep_graph_nodes_search. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `direction` (`string`, optional): The edge orientation to follow from the node: in, out, or both. Example: in. One of: `in`, `out`, `both`.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `order` (`string`, optional): Sort direction: asc or desc. Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort field. Example: created_at. One of: `uuid`, `created_at`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.

Example input: `{"graph_uuid":"<graph_uuid>","node_uuid":"<node_uuid>"}`

### `zep_nodes_list`

List Nodes · Read-only

List a graph's nodes (entities), a page at a time, optionally filtered by label. Returns node objects with name, labels and summary. Use zep_graph_nodes_search for a text query. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `order` (`string`, optional): Sort direction: asc or desc (default desc). Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort key: uuid (default), created_at, or degree. Example: created_at. One of: `uuid`, `created_at`, `degree`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_observations_list`

List Observations · Read-only

List the observations Zep derived from a graph's data, a page at a time. Returns observation objects. Use zep_graph_observations_search for a text query. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `order` (`string`, optional): Sort direction: asc or desc (default desc). Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort key: uuid (default) or created_at. Example: created_at. One of: `uuid`, `created_at`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_project_get`

Get Project · Read-only

Get the Zep project the API key belongs to. Returns the project's UUID, name and settings. Use it to test the connection.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `zep_project_instructions_get`

Get Project Instructions · Read-only

Get the project-wide custom extraction instructions. Returns the instructions. Use zep_graph_instructions_get for one graph.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `zep_project_ontology_get`

Get Project Ontology · Read-only

Get the project-wide default ontology (entity and edge types). Returns the ontology. Use zep_graph_ontology_get for one graph.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `zep_task_get`

Get Task · Read-only

Get the status of a background task such as a graph clone. Returns the task with status and progress. Get the task UUID from the call that started it, such as zep_graph_clone.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `task_uuid` (`string`, required): UUID of the task returned by the call that started it. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.

Example input: `{"task_uuid":"<task_uuid>"}`

### `zep_thread_context_get`

Get Thread Context · Read-only

Get the context block for a thread: the relevant facts, entities and summaries from the user's graph for the latest messages, ready to put in an LLM prompt. Returns the context string. Use zep_graph_context_get for a custom query against any graph. Get thread UUIDs from zep_threads_list or zep_thread_lookup.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `thread_uuid` (`string`, required): Zep UUID of the thread. Get it from zep_threads_list or zep_thread_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `template_uuid` (`string`, optional): Context template UUID. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.

Example input: `{"thread_uuid":"<thread_uuid>"}`

### `zep_thread_episodes_list`

List Thread Episodes · Read-only

List the graph episodes created from a thread's messages, with their processing status. Returns episode objects. Use zep_episode_get for one episode. Get thread UUIDs from zep_threads_list or zep_thread_lookup.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `thread_uuid` (`string`, required): Zep UUID of the thread. Get it from zep_threads_list or zep_thread_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.

Example input: `{"thread_uuid":"<thread_uuid>"}`

### `zep_thread_get`

Get Thread · Read-only

Get one thread by Zep UUID. Returns the thread with its user and graph. Get thread UUIDs from zep_threads_list or zep_thread_lookup.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `thread_uuid` (`string`, required): Zep UUID of the thread. Get it from zep_threads_list or zep_thread_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"thread_uuid":"<thread_uuid>"}`

### `zep_thread_lookup`

Look Up Thread by Developer ID · Read-only

Find a thread by the developer ID (your own identifier) you gave it. Returns the matching thread. Use zep_thread_get with a Zep UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_id` (`string`, optional): The developer-assigned graph ID to resolve to a UUID. Mutually exclusive with user_id and thread_id. Example: product-docs.
- `thread_id` (`string`, optional): The developer-assigned thread ID to resolve to a UUID. Mutually exclusive with user_id and graph_id. Example: support-chat-2026-01-15.
- `user_id` (`string`, optional): The developer-assigned user ID to resolve to a UUID. Mutually exclusive with thread_id and graph_id. Example: auth0|64f1c2.

Example input: `{}`

### `zep_thread_message_get`

Get Thread Message · Read-only

Get one message from a thread. Returns the message object. Get message UUIDs from zep_thread_messages_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `message_uuid` (`string`, required): UUID of the message. Get it from zep_thread_messages_list. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `thread_uuid` (`string`, required): Zep UUID of the thread. Get it from zep_threads_list or zep_thread_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"message_uuid":"<message_uuid>","thread_uuid":"<thread_uuid>"}`

### `zep_thread_messages_list`

List Thread Messages · Read-only

List the messages in a thread, a page at a time. Returns message objects with role, content and time. Get thread UUIDs from zep_threads_list or zep_thread_lookup.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `thread_uuid` (`string`, required): Zep UUID of the thread. Get it from zep_threads_list or zep_thread_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `order` (`string`, optional): Sort direction: asc or desc. Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort field. Example: created_at. One of: `created_at`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.

Example input: `{"thread_uuid":"<thread_uuid>"}`

### `zep_thread_summaries_list`

List Thread Summaries · Read-only

List the thread summaries stored in a graph, a page at a time. Returns summary objects. Use zep_graph_thread_summaries_search for a text query. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `filters` (`object`, optional): Filters that narrow the search, as a JSON object (for example node_labels, edge_types, created_at ranges). Example: {"node_labels": ["Person"]}
- `order` (`string`, optional): Sort direction: asc or desc (default desc). Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort key: uuid (default) or created_at. Example: created_at. One of: `uuid`, `created_at`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_thread_summary_get`

Get Thread Summary · Read-only

Get the running summary Zep keeps of a thread. Returns the summary text. Get thread UUIDs from zep_threads_list or zep_thread_lookup.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `thread_uuid` (`string`, required): Zep UUID of the thread. Get it from zep_threads_list or zep_thread_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"thread_uuid":"<thread_uuid>"}`

### `zep_threads_list`

List Threads · Read-only

List the threads in the project, a page at a time, newest or oldest first. Returns thread objects with their user. Use zep_thread_lookup to find a thread by your own ID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `order` (`string`, optional): asc or desc. Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort field. Example: created_at. One of: `created_at`, `updated_at`, `uuid`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.
- `user_uuid` (`string`, optional): Zep UUID of the user. Get it from zep_users_list or zep_user_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{}`

### `zep_user_get`

Get User · Read-only

Get one user by Zep UUID with name, email and metadata. Returns the user object. Get user UUIDs from zep_users_list or zep_user_lookup.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `user_uuid` (`string`, required): Zep UUID of the user. Get it from zep_users_list or zep_user_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"user_uuid":"<user_uuid>"}`

### `zep_user_get_by_email`

Get User by Email · Read-only

Find a user by email address. Returns the matching user. Use zep_user_lookup for your own IDs.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `email` (`string`, required): The user's email address. Matching ignores case and leading and trailing whitespace. Example: jane.doe@example.com.

Example input: `{"email":"<email>"}`

### `zep_user_lookup`

Look Up User by Developer ID · Read-only

Find a user by the developer ID (your own identifier) you gave it. Returns the matching user. Use zep_user_get when you already have the Zep UUID.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_id` (`string`, optional): The developer-assigned graph ID to resolve to a UUID. Mutually exclusive with user_id and thread_id. Example: product-docs.
- `thread_id` (`string`, optional): The developer-assigned thread ID to resolve to a UUID. Mutually exclusive with user_id and graph_id. Example: support-chat-2026-01-15.
- `user_id` (`string`, optional): The developer-assigned user ID to resolve to a UUID. Mutually exclusive with thread_id and graph_id. Example: auth0|64f1c2.

Example input: `{}`

### `zep_user_node_get`

Get User Node · Read-only

Get the node that represents the user in their knowledge graph, with its summary. Returns the node. Use zep_nodes_list for other nodes. Get user UUIDs from zep_users_list or zep_user_lookup.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `user_uuid` (`string`, required): Zep UUID of the user. Get it from zep_users_list or zep_user_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"user_uuid":"<user_uuid>"}`

### `zep_user_summary_instructions_get`

Get User Summary Instructions · Read-only

Get the custom instructions Zep follows when summarising this user. Returns the instructions. Get user UUIDs from zep_users_list or zep_user_lookup.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `user_uuid` (`string`, required): Zep UUID of the user. Get it from zep_users_list or zep_user_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"user_uuid":"<user_uuid>"}`

### `zep_users_list`

List Users · Read-only

List the users in the project, a page at a time. Returns user objects with their UUIDs. Use zep_user_lookup to find users by your own IDs and zep_user_get_by_email by email.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `cursor` (`string`, optional): Opaque page cursor from the previous response, to get the next page. Example: eyJvZmZzZXQiOjUwfQ.
- `order` (`string`, optional): asc or desc. Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Sort field. Example: created_at. One of: `created_at`, `user_id`, `email`, `uuid`.
- `page_size` (`integer`, optional): Maximum number of items in one page. Defaults to 50. Example: 50.
- `search` (`string`, optional): Filters results to users whose user ID, email, or name contains this text. Example: quarterly report.

Example input: `{}`

### `zep_edge_update`

Update Edge · Write

Update an edge's fact, name, attributes or validity times. Returns the updated edge. Get UUIDs from zep_edges_list.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `edge_uuid` (`string`, required): UUID of the edge. Get it from zep_edges_list or zep_graph_edges_search. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `attributes` (`object`, optional): Additional attributes to merge onto the edge; a key set to null is removed. Example: {"price": "$49"}
- `fact` (`string`, optional): The fact text describing the relationship between the source and target nodes. Example: Jane is on the Pro plan.

Example input: `{"edge_uuid":"<edge_uuid>","graph_uuid":"<graph_uuid>"}`

### `zep_edges_add`

Add Edges · Write

Add facts as edges between existing nodes of a graph. Returns the created edges. Get node UUIDs from zep_nodes_list. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `edges` (`array`, required): Edges to add, each with the source and target node UUIDs, a name (relation type) and the fact text, plus optional validity times. Example: [{"source_node_uuid": "3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d", "target_node_uuid": "8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e", "name": "SUBSCRIBED_TO", "fact": "J...
- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `deduplicate` (`boolean`, optional): When true, Zep compares each edge with graph edges and can merge a duplicate or invalidate a contradicted edge. This adds an LLM call per edge. The default is false. Example: true.

Example input: `{"edges":[],"graph_uuid":"<graph_uuid>"}`

### `zep_episode_add`

Add Episode · Write

Add data to a graph as an episode: text, a JSON document or a message, which Zep processes into nodes and edges in the background. Returns the episode with its processing status. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `data` (`string`, required): The episode content to add to the graph. Example: Jane upgraded to the Pro plan on 2026-01-15.
- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `document_id` (`string`, optional): Groups this episode as a chunk of a document on the graph. Example: pricing-faq.
- `metadata` (`object`, optional): Key-value metadata to store, as a JSON object. Example: {"plan": "pro"}
- `reference_time` (`string`, optional): The episode's reference time, used for temporal reasoning rather than ingestion time. Defaults to the ingestion time. Example: 2026-01-15T09:30:00Z.
- `source_description` (`string`, optional): A description of the source of this episode. Example: Support chat transcript.
- `strict_ontology` (`boolean`, optional): When true, prevents extraction of generic entity nodes that do not match the configured ontology. Example: true.
- `type` (`string`, optional): The data format of the episode: text, json, or message. Defaults to text. Example: text. One of: `text`, `json`, `message`.

Example input: `{"data":"<data>","graph_uuid":"<graph_uuid>"}`

### `zep_graph_clone`

Clone Graph · Write

Copy a graph (or a user's graph) into a new graph or user, for testing or templates. Returns the new graph or user and a task to poll with zep_task_get. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_graph_create`

Create Graph · Write

Create a standalone knowledge graph (not tied to a user) with a developer ID, name and description. Returns the graph with its Zep UUID. Add data with zep_episode_add.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `content_policy` (`object`, optional): Content policy additions for the graph. The graph binds the current project content policy plus these additions, and the binding does not change after creation. Example: {"rules": []}
- `description` (`string`, optional): A description of the graph. Example: Knowledge graph of the product documentation.
- `name` (`string`, optional): A display name for the graph. Example: Product docs.
- `time_zone` (`string`, optional): The graph's IANA time zone. Example: America/New_York.

Example input: `{}`

### `zep_graph_instructions_set`

Set Graph Instructions · Write

Replace a graph's custom extraction instructions, which tell Zep what to focus on when building the graph. Returns the saved instructions. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `instructions` (`array`, optional): Instructions to apply, each an object with a short name and the instruction text. This replaces the existing instructions at this scope. Example: [{"name": "billing-focus", "text": "Focus on the customer's plan, billing issues and product feedback."}]

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_graph_ontology_set`

Set Graph Ontology · Write

Replace a graph's custom entity and edge types (ontology), which guide what Zep extracts. Returns the saved ontology. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `edge_types` (`array`, optional): Edge (fact) types to extract, each with a name, a description, optional properties and the source and target entity types it connects. Example: [{"name": "SUBSCRIBED_TO", "description": "A customer subscribed to a plan", "source_targets": [{"source": "Person", "target": "Plan"}]}]
- `entity_types` (`array`, optional): Entity types to extract, each with a name, a description and optional typed properties (text, int, float, boolean, datetime). Example: [{"name": "Plan", "description": "A subscription plan", "properties": [{"name": "price", "type": "text", "description": "Monthly price"}]}]

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_graph_update`

Update Graph · Write

Update a graph's name, description or developer ID. Returns the updated graph. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `description` (`string`, optional): A description of the graph. Example: Knowledge graph of the product documentation.
- `name` (`string`, optional): The graph's display name. Example: Product docs.
- `time_zone` (`string`, optional): The graph's IANA time zone. Example: America/New_York.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_node_update`

Update Node · Write

Update a node's name, labels, summary or attributes. Returns the updated node. Get UUIDs from zep_nodes_list.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `node_uuid` (`string`, required): UUID of the node. Get it from zep_nodes_list or zep_graph_nodes_search. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `attributes` (`object`, optional): Additional attributes to merge onto the node; a key set to null is removed. Example: {"price": "$49"}
- `name` (`string`, optional): The node's name. Example: Product docs.
- `summary` (`string`, optional): A summary of the node. Example: Jane is a Pro customer interested in SSO.

Example input: `{"graph_uuid":"<graph_uuid>","node_uuid":"<node_uuid>"}`

### `zep_nodes_add`

Add Nodes · Write

Add entities as nodes to a graph, with labels, summary and attributes. Returns the created nodes. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `nodes` (`array`, required): Nodes to add, each with a name and optionally labels (entity types), a summary and attributes. Example: [{"name": "Pro plan", "labels": ["Plan"], "summary": "The paid Pro subscription tier"}]

Example input: `{"graph_uuid":"<graph_uuid>","nodes":[]}`

### `zep_project_instructions_set`

Set Project Instructions · Write

Replace the project-wide custom extraction instructions applied to every graph. Returns the saved instructions. Use zep_graph_instructions_set for one graph.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `instructions` (`array`, optional): Instructions to apply, each an object with a short name and the instruction text. This replaces the existing instructions at this scope. Example: [{"name": "billing-focus", "text": "Focus on the customer's plan, billing issues and product feedback."}]

Example input: `{}`

### `zep_thread_create`

Create Thread · Write

Create a conversation thread for a user. Returns the thread with its Zep UUID. Add messages with zep_thread_messages_add. Get user UUIDs from zep_users_list or zep_user_lookup.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `user_uuid` (`string`, required): Zep UUID of the user. Get it from zep_users_list or zep_user_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"user_uuid":"<user_uuid>"}`

### `zep_thread_messages_add`

Add Thread Messages · Write

Add one or more chat messages to a thread; Zep stores them and extracts facts into the user's graph in the background. Returns the stored messages and, with return_context, the updated context block. Get thread UUIDs from zep_threads_list or zep_thread_lookup.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `messages` (`array`, required): Messages to add, each with a role (user, assistant, system, tool, ...), the content and optionally the speaker's name, created_at and metadata. Example: [{"role": "user", "name": "Jane", "content": "I just upgraded to Pro."}]
- `thread_uuid` (`string`, required): Zep UUID of the thread. Get it from zep_threads_list or zep_thread_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `ignore_roles` (`array`, optional): Message roles to skip during graph extraction; the messages are still stored. Example: ["assistant"]
- `return_context` (`boolean`, optional): When true, returns the context block for the thread's most recent messages. Example: true.
- `strict_ontology` (`boolean`, optional): When true, prevents extraction of generic entity nodes that do not match the configured ontology. Example: true.

Example input: `{"messages":[{"role":"user","name":"Jane","content":"I just upgraded to Pro."}],"thread_uuid":"<thread_uuid>"}`

### `zep_user_create`

Create User · Write

Create a user in the Zep project, with optional name, email, external ID and metadata; Zep creates the user's knowledge graph. Returns the user with its Zep UUID. Use zep_thread_create next to start a conversation thread for the user.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `content_policy` (`object`, optional): Content policy additions for the user's graph. The graph binds the current project content policy plus these additions, and the binding does not change after creation. Example: {"rules": []}
- `disable_default_ontology` (`boolean`, optional): When true, disables the default ontology for the user's graph. Example: true.
- `email` (`string`, optional): The email address of the user. Example: jane.doe@example.com.
- `first_name` (`string`, optional): The user's first name. Example: Jane.
- `last_name` (`string`, optional): The user's last name. Example: Doe.
- `metadata` (`object`, optional): Key-value metadata to store, as a JSON object. Example: {"plan": "pro"}
- `time_zone` (`string`, optional): The user's IANA time zone. Example: America/New_York.
- `user_id` (`string`, optional): An optional external identity identifier, such as an identity-provider subject claim. Zep uses it to match an identity-provider sign-in to this user. It is not an address; use uuid to address the user in v4. It is unique within the project, can contain at most 500 characters, and cannot be changed after creation. Example: auth0|64f1c2.

Example input: `{}`

### `zep_user_summary_instructions_set`

Set User Summary Instructions · Write

Replace the custom instructions Zep follows when summarising this user. Returns the saved instructions. Get user UUIDs from zep_users_list or zep_user_lookup.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `user_uuid` (`string`, required): Zep UUID of the user. Get it from zep_users_list or zep_user_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `instructions` (`array`, optional): Instructions to apply, each an object with a short name and the instruction text. This replaces the existing instructions at this scope. Example: [{"name": "billing-focus", "text": "Focus on the customer's plan, billing issues and product feedback."}]

Example input: `{"user_uuid":"<user_uuid>"}`

### `zep_user_update`

Update User · Write

Update a user's name, email, time zone or metadata. Returns the updated user. Get user UUIDs from zep_users_list or zep_user_lookup.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `user_uuid` (`string`, required): Zep UUID of the user. Get it from zep_users_list or zep_user_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `disable_default_ontology` (`boolean`, optional): When true, disables the default ontology for the user's graph. Example: true.
- `email` (`string`, optional): The email address of the user. Example: jane.doe@example.com.
- `first_name` (`string`, optional): The user's first name. Example: Jane.
- `last_name` (`string`, optional): The user's last name. Example: Doe.
- `metadata` (`object`, optional): Key-value metadata to store, as a JSON object. Example: {"plan": "pro"}
- `time_zone` (`string`, optional): The user's IANA time zone. Example: America/New_York.

Example input: `{"user_uuid":"<user_uuid>"}`

### `zep_edge_delete`

Delete Edge · Destructive

Delete an edge (fact) from a graph. Returns an empty response. Get UUIDs from zep_edges_list.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `edge_uuid` (`string`, required): UUID of the edge. Get it from zep_edges_list or zep_graph_edges_search. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"edge_uuid":"<edge_uuid>","graph_uuid":"<graph_uuid>"}`

### `zep_episode_delete`

Delete Episode · Destructive

Delete an episode and the edges that only it supports. Returns an empty response. Get UUIDs from zep_episodes_list.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `episode_uuid` (`string`, required): UUID of the episode. Get it from zep_episodes_list. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.
- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"episode_uuid":"<episode_uuid>","graph_uuid":"<graph_uuid>"}`

### `zep_graph_delete`

Delete Graph · Destructive

Delete a graph with all its nodes, edges and episodes. Returns an empty response. Get graph UUIDs from zep_graphs_list; for a user's graph pass the user's UUID.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"graph_uuid":"<graph_uuid>"}`

### `zep_node_delete`

Delete Node · Destructive

Delete a node and its edges from a graph. Returns an empty response. Get UUIDs from zep_nodes_list.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `graph_uuid` (`string`, required): Zep UUID of the graph. For a user's graph you can also pass the user's UUID. Get graph UUIDs from zep_graphs_list. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.
- `node_uuid` (`string`, required): UUID of the node. Get it from zep_nodes_list or zep_graph_nodes_search. Example: 8a1c2d3e-4f5a-4b6c-8d7e-9f0a1b2c3d4e.

Example input: `{"graph_uuid":"<graph_uuid>","node_uuid":"<node_uuid>"}`

### `zep_thread_delete`

Delete Thread · Destructive

Delete a thread and its messages. Facts already extracted into the user's graph stay. Returns an empty response. Get thread UUIDs from zep_threads_list or zep_thread_lookup.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `thread_uuid` (`string`, required): Zep UUID of the thread. Get it from zep_threads_list or zep_thread_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"thread_uuid":"<thread_uuid>"}`

### `zep_user_delete`

Delete User · Destructive

Delete a user with their threads and knowledge graph. Returns an empty response. Get user UUIDs from zep_users_list or zep_user_lookup.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `user_uuid` (`string`, required): Zep UUID of the user. Get it from zep_users_list or zep_user_lookup. Example: 3f2b8c1e-7d4a-4e5b-9c6d-2a1b0e9f8c7d.

Example input: `{"user_uuid":"<user_uuid>"}`

## Related

Other AI connectors ([all 216](/agentkit/connectors/?category=ai)):

- [Chorus](/agentkit/connectors/chorus/): Scalekit connector, Username and password, 9 tools
- [Gong](/agentkit/connectors/gong/): Scalekit connector, OAuth, 69 tools
- [Jiminny](/agentkit/connectors/jiminny/): Scalekit connector, Access token, 27 tools


---

## More Scalekit documentation

| Resource | What it contains | When to use it |
|----------|-----------------|----------------|
| [/llms.txt](/llms.txt) | Structured index with routing hints per product area | Start here — find which documentation set covers your topic before loading full content |
| [/llms-full.txt](/llms-full.txt) | Complete documentation for all Scalekit products in one file | Use when you need exhaustive context across multiple products or when the topic spans several areas |
| [sitemap-0.xml](https://docs.scalekit.com/sitemap-0.xml) | Full URL list of every documentation page | Use to discover specific page URLs you can fetch for targeted, page-level answers |
