> **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 Langfuse

The Langfuse connector lets your AI agent work with Langfuse. Scalekit handles authentication for every call, so your agent never handles credentials. It comes with 58 tools.

**Authentication:** Public Key + Secret Key
**Categories:** AI, Monitoring
**Tools:** 58: 32 read, 16 write, 10 destructive
**Users sign in with:** Username and password
**Built by:** Scalekit connector
**Try it:** [Playground in the Scalekit dashboard](https://app.scalekit.com/ws/signup?sk_intent=playground&provider=LANGFUSE)

## 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 Langfuse connection

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

## 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: 'langfuse_annotation_queue_get',
  toolInput: {
    queue_id: '<queue_id>',
  },
  connector: 'langfuse',
  identifier: 'user_123',
})
```

**Python**

```python
result = actions.execute_tool(
    tool_name="langfuse_annotation_queue_get",
    tool_input={
        "queue_id": "<queue_id>",
    },
    connection_name="langfuse",
    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.

### `langfuse_annotation_queue_get`

Get Annotation Queue · Read-only

Get one annotation queue.
Returns its name, description and score configs.

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

Inputs:

- `queue_id` (`string`, required): ID of the annotation queue. Get it from langfuse_annotation_queues_list. Example: cm1q2w3e4r5t6y7u8i9o0p.

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

### `langfuse_annotation_queue_item_get`

Get Annotation Queue Item · Read-only

Get one item of an annotation queue.
Returns the item's object, status and completion info.

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

Inputs:

- `item_id` (`string`, required): ID of the queue item. Example: cm2a3s4d5f6g7h8j9k0l1z.
- `queue_id` (`string`, required): ID of the annotation queue. Get it from langfuse_annotation_queues_list. Example: cm1q2w3e4r5t6y7u8i9o0p.

Example input: `{"item_id":"<item_id>","queue_id":"<queue_id>"}`

### `langfuse_annotation_queue_items_list`

List Annotation Queue Items · Read-only

List the items in an annotation queue, optionally by status, with paging.
Returns each item's ID, object (trace, observation or session), status and timestamps.

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

Inputs:

- `queue_id` (`string`, required): ID of the annotation queue. Get it from langfuse_annotation_queues_list. Example: cm1q2w3e4r5t6y7u8i9o0p.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.
- `status` (`string`, optional): Filter by status. Example: PENDING. One of: `PENDING`, `COMPLETED`.

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

### `langfuse_annotation_queues_list`

List Annotation Queues · Read-only

List annotation queues, with paging.
Returns each queue's ID, name, description and score configs.

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

Inputs:

- `limit` (`integer`, optional): Items per page. Example: 50.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.

Example input: `{}`

### `langfuse_comment_get`

Get Comment · Read-only

Get one comment.
Returns its object, content, author and time.

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

Inputs:

- `comment_id` (`string`, required): The unique langfuse identifier of a comment. Example: cm3x4c5v6b7n8m9q0w1e2r.

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

### `langfuse_comments_list`

List Comments · Read-only

List comments, optionally on one object, with paging.
Returns each comment's ID, object, content, author and time.

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

Inputs:

- `author_user_id` (`string`, optional): Filter comments by author user id. Example: user-1.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `object_id` (`string`, optional): Filter comments by object id. If objectType is not provided, an error will be thrown. Example: 4bf92f3577b34da6a3ce929d0e0e4736.
- `object_type` (`string`, optional): Filter comments by object type (trace, observation, session, prompt). Example: TRACE.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.

Example input: `{}`

### `langfuse_dataset_get`

Get Dataset · Read-only

Get one dataset by name.
Returns the dataset's ID, description and metadata.

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

Inputs:

- `dataset_name` (`string`, required): Dataset Name. Example: golden-qa.

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

### `langfuse_dataset_item_get`

Get Dataset Item · Read-only

Get one dataset item.
Returns its input, expected output, metadata, status and source.

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

Inputs:

- `dataset_item_id` (`string`, required): ID of the dataset item. Example: item-001.

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

### `langfuse_dataset_items_list`

List Dataset Items · Read-only

List dataset items, optionally for one dataset or source trace, with paging.
Returns each item's ID, input, expected output, metadata and status.

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

Inputs:

- `dataset_name` (`string`, optional): Dataset Name. Example: golden-qa.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.
- `source_observation_id` (`string`, optional): Source Observation ID. Example: obs-1.
- `source_trace_id` (`string`, optional): Source Trace ID. Example: 4bf92f3577b34da6a3ce929d0e0e4736.
- `version` (`string`, optional): ISO 8601 timestamp (RFC 3339, Section 5.6) in UTC (e.g., "2026-01-21T14:35:42Z"). If provided, returns state of dataset at this timestamp. If not provided, returns the latest version. Requires datasetName to be specified. Example: 2026-05-01T00:00:00Z.

Example input: `{}`

### `langfuse_datasets_list`

List Datasets · Read-only

List datasets in the project, with paging.
Returns each dataset's ID, name, description and metadata.

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

Inputs:

- `limit` (`integer`, optional): Items per page. Example: 50.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.

Example input: `{}`

### `langfuse_evaluation_rule_get`

Get Evaluation Rule · Read-only

Get one evaluation rule.
Returns its name, target, filter, sampling and evaluator assignments.

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

Inputs:

- `evaluation_rule_id` (`string`, required): Stable evaluation-rule identifier returned by the evaluation-rule endpoints. Example: cm6n7m8q9w0e1r2t3y4u5i.

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

### `langfuse_evaluation_rules_list`

List Evaluation Rules · Read-only

List evaluation rules, which run evaluators on new traces or dataset runs.
Returns each rule's ID, name, enabled flag, sampling, filter and evaluators.

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

Inputs:

- `cursor` (`string`, optional): Cursor from the previous response's meta, for the next page. Example: eyJ0IjoxfQ.
- `limit` (`integer`, optional): Items per page. Example: 50.

Example input: `{}`

### `langfuse_evaluator_get`

Get Evaluator · Read-only

Get one evaluator.
Returns its name, type, prompt or code, model config and output definition.

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

Inputs:

- `evaluator_id` (`string`, required): Stable evaluator identifier returned by the evaluator endpoints. Example: cm7o8p9a0s1d2f3g4h5j6k.

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

### `langfuse_evaluator_versions_list`

List Evaluator Versions · Read-only

List an evaluator's version history, newest first.
Returns each version's number and definition.

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

Inputs:

- `evaluator_id` (`string`, required): Stable evaluator identifier returned by the evaluator endpoints. Example: cm7o8p9a0s1d2f3g4h5j6k.
- `cursor` (`string`, optional): Cursor from the previous response's meta, for the next page. Example: eyJ0IjoxfQ.
- `limit` (`integer`, optional): Items per page. Example: 50.

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

### `langfuse_evaluators_list`

List Evaluators · Read-only

List evaluators (LLM-as-a-judge and code evaluators) in the project.
Returns each evaluator's ID, name, type and latest version.

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

Inputs:

- `cursor` (`string`, optional): Cursor from the previous response's meta, for the next page. Example: eyJ0IjoxfQ.
- `limit` (`integer`, optional): Items per page. Example: 50.

Example input: `{}`

### `langfuse_experiment_items_list`

List Experiment Items · Read-only

List the items of experiments with cursor paging.
Returns each item's experiment, dataset item, trace and observation.

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

Inputs:

- `from_start_time` (`string`, required): Required. Retrieve only experiment items started on or after this datetime. A lower bound is required so the query stays fast on large projects; pair it with an upper bound to select a specific window. Example: 2026-01-01T00:00:00Z.
- `cursor` (`string`, optional): Cursor from the previous response's meta, for the next page. Example: eyJ0IjoxfQ.
- `dataset_id` (`string`, optional): Comma-separated list of dataset IDs. Example: 1234567890.
- `experiment_id` (`string`, optional): Comma-separated list of experiment IDs. Example: 1234567890.
- `experiment_item_id` (`string`, optional): Comma-separated list of experiment item IDs. Example: 1234567890.
- `experiment_name` (`string`, optional): Comma-separated list of experiment names. Example: example.
- `fields` (`string`, optional): Comma-separated list of field groups to include. Available groups: `core`, `dataset`, `io`, `metadata`, `itemMetadata`, `experimentMetadata`, `scores`. If omitted, `core,dataset` is returned. Example: example.
- `filter` (`string`, optional): JSON string containing an array of structured filter conditions. Supported columns are `experimentId`, `experimentName`, `experimentItemId`, and `datasetId`. Example: example.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `score_limit` (`integer`, optional): Number of scores to return per experiment item when `fields=scores` is requested. Maximum 50, default 50. Example: 1.
- `to_start_time` (`string`, optional): Retrieve only experiment items started before this datetime. Example: 2026-01-01T00:00:00Z.

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

### `langfuse_experiments_list`

List Experiments · Read-only

List experiments (dataset runs) with cursor paging.
Returns each experiment's ID, name, dataset, metadata and timestamps.

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

Inputs:

- `from_start_time` (`string`, required): Required. Retrieve only experiments with events on or after this datetime. A lower bound is required so the query stays fast on large projects; pair it with an upper bound to select a specific window. Example: 2026-01-01T00:00:00Z.
- `cursor` (`string`, optional): Cursor from the previous response's meta, for the next page. Example: eyJ0IjoxfQ.
- `dataset_id` (`string`, optional): Comma-separated list of dataset IDs. Example: 1234567890.
- `fields` (`string`, optional): Comma-separated list of field groups to include. Available groups: `core`, `metadata`, `scores`. If omitted, `core` is returned. Example: example.
- `filter` (`string`, optional): JSON string containing an array of structured filter conditions. Supported columns are `id`, `name`, and `datasetId`. Example: example.
- `id` (`string`, optional): Comma-separated list of experiment IDs. Example: 1234567890.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `name` (`string`, optional): Comma-separated list of experiment names. Example: support-answer.
- `score_limit` (`integer`, optional): Number of scores to return per experiment when `fields=scores` is requested. Maximum 50, default 50. Example: 1.
- `to_start_time` (`string`, optional): Retrieve only experiments before this datetime. Example: 2026-01-01T00:00:00Z.

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

### `langfuse_health_check`

Check API Health · Read-only

Check that the Langfuse API and its database are up.
Returns the status and version.
Run it to confirm the host and keys.

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

Inputs: none.

Example input: `{}`

### `langfuse_llm_connections_list`

List LLM Connections · Read-only

List the project's LLM provider connections used by evaluators and the playground.
Returns each connection's ID, provider, adapter and models (no secret keys).

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

Inputs:

- `limit` (`integer`, optional): Items per page. Example: 50.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.

Example input: `{}`

### `langfuse_media_get`

Get Media · Read-only

Get a media record attached to a trace or observation.
Returns its content type, size and a temporary download URL.

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

Inputs:

- `media_id` (`string`, required): The unique langfuse identifier of a media record. Example: cm8l9z0x1c2v3b4n5m6q7w.

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

### `langfuse_metrics_query`

Query Metrics · Read-only

Run an aggregate metrics query over traces, observations or scores: counts, latency, tokens and cost, grouped by dimensions and time.
Returns the result rows.
Use it for usage and cost reports.

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

Inputs:

- `query` (`string`, required): The query as a JSON string with view (traces, observations, scores-numeric, scores-categorical), metrics, dimensions, filters, timeDimension, fromTimestamp and toTimestamp. Example: {"view": "observations", "metrics": [{"measure": "totalCost", "aggregation": "sum"}], "dimensions": [{"field": "providedModelName"}], "filters": [], "fromTimestamp": "2026-05-01T00:00:00Z", "toTimestamp": "2026-05-31T23:59:59Z"}.

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

### `langfuse_model_get`

Get Model · Read-only

Get one model definition.
Returns its name, match pattern, unit, prices and tokenizer.

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

Inputs:

- `model_id` (`string`, required): ID of the model. Get it from langfuse_models_list. Example: cm0q1w2e3r4t5y6u7i8o9p.

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

### `langfuse_models_list`

List Models · Read-only

List model definitions used for token and cost tracking, with paging.
Returns each model's ID, name, match pattern, prices and tokenizer.

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

Inputs:

- `limit` (`integer`, optional): Items per page. Example: 50.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.

Example input: `{}`

### `langfuse_observations_list`

List Observations · Read-only

List observations (spans, generations and events) with filters by trace, session, user, name, type, level, environment and start time, with cursor paging.
Returns each observation's IDs, type, name, timing, model, usage, cost, level and, if requested, input and output.
Use it to inspect traces and LLM calls.

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

Inputs:

- `cursor` (`string`, optional): Cursor from the previous response's meta, for the next page. Example: eyJ0IjoxfQ.
- `environment` (`array`, optional): Optional filter for observations where the environment is one of the provided values. Example: ["production"].
- `expand_metadata` (`string`, optional): Comma-separated list of metadata keys to return non-truncated. By default, metadata values over 200 characters are truncated. Use this parameter to retrieve full values for specific keys. Example: "key1,key2".
- `fields` (`string`, optional): Comma-separated list of field groups to include in the response. Available groups: core, basic, time, io, metadata, model, usage, prompt, metrics, trace_context. If not specified, `core` and `basic` field groups are returned. Example: "basic,usage,model".
- `filter` (`string`, optional): JSON string containing an array of filter conditions. A condition on the same field as a query parameter filter (userId, name, type, level, environment, fromStartTime, .) replaces that parameter. Query parameters for other fields still apply. A `startTime` condition replaces both `fromStartTime` and `toStartTime`. ## Filter Structure Each filter condition has the following structure: ```json [ { "type": string, // Required. One of: "datetime", "string", "number", "stringOptions", "arrayOptions", "stringObject", "boolean", "null" "column": string, // Required. Column to filter on (see available columns below) "operator": string, // Required. Operator based on type: // - datetime: ">", "<", ">=", "<=" // - string: "=", "contains", "does not contain", "starts with", "ends with", "is not empty", "matches" (input/output only) // - stringOptions: "any of", "none of" (works on any string column except input/output) // - arrayOptions: "any of", "none of", "all of" // - number: "=", ">", "<", ">=", "<=" // - stringObject: "=", "contains", "does not contain", "starts with", "ends with", "matches", "is set", "is not set" // - boolean: "=", "<>" // - null: "is null", "is not null" "value": any, // Required. Type depends on filter type: ISO 8601 string for datetime, JSON number for number (not a numeric string), array of strings for stringOptions/arrayOptions, boolean for boolean, and "" (empty string) for null "key": string // Required only for stringObject (the metadata key to filter on) } ] ``` ## Available Columns ### Core Observation Fields - `id` (string) - Observation ID - `type` (string) - Observation type (SPAN, GENERATION, EVENT) - `name` (string) - Observation name - `traceId` (string) - Associated trace ID - `startTime` (datetime) - Observation start time - `endTime` (datetime) - Observation end time - `environment` (string) - Environment tag - `level` (string) - Log level (DEBUG, DEFAULT, WARNING, ERROR) - `statusMessage` (string) - Status message - `version` (string) - Version tag - `userId` (string) - User ID - `sessionId` (string) - Session ID - `isRootObservation` (boolean) - Whether the observation is a logical root. Observations marked as app roots by the SDK may retain a non-null parentObservationId. ### Trace-Related Fields - `traceName` (string) - Name of the parent trace - `traceTags` (arrayOptions) - Tags from the parent trace - `tags` (arrayOptions) - Alias for traceTags ### Performance Metrics - `latency` (number) - Latency in seconds (calculated: end_time - start_time) - `timeToFirstToken` (number) - Time to first token in seconds - `tokensPerSecond` (number) - Output tokens per second ### Token Usage - `inputTokens` (number) - Number of input tokens - `outputTokens` (number) - Number of output tokens - `totalTokens` (number) - Total tokens (alias: `tokens`) ### Cost Metrics - `inputCost` (number) - Input cost in USD - `outputCost` (number) - Output cost in USD - `totalCost` (number) - Total cost in USD ### Model Information - `model` (string) - Provided model name (alias: `providedModelName`) - `promptName` (string) - Associated prompt name - `promptVersion` (number) - Associated prompt version ### Structured Data - `input` (string) - Observation input. Supports accelerated indexed literal search with the `matches` operator. - `output` (string) - Observation output. Supports accelerated indexed literal search with the `matches` operator. - `metadata` (stringObject) - Metadata key-value pairs. Use the `key` parameter to filter on a specific metadata key. Metadata values are stored as strings, so match numbers and booleans by their string form with `=` (for example `"value": "true"` or `"value": "42"`); numeric comparisons such as `>` are not supported on metadata. The `matches` operator is only supported for `input`, `output`, and stringObject `metadata` filters. It performs indexed literal search with token-boundary pruning using the events table text indexes. Case sensitivity differs by target: `input` and `output` matches are case-insensitive, while metadata value matches are case-sensitive. Unlike SQL `LIKE`, `%` and `_` are treated as literal characters. Use `contains` for legacy substring semantics where the API allows it. Any v2 `input` or `output` filter must be accompanied by at least one `=` or `matches` filter on `input` or `output`; standalone `contains`, `starts with`, `ends with`, and `does not contain` filters on these columns are rejected. ## Filter Examples ```json [ { "type": "string", "column": "type", "operator": "=", "value": "GENERATION" }, { "type": "number", "column": "latency", "operator": ">=", "value": 2.5 }, { "type": "stringObject", "column": "metadata", "key": "environment", "operator": "=", "value": "production" }, { "type": "string", "column": "output", "operator": "matches", "value": "needle" }, { "type": "boolean", "column": "isRootObservation", "operator": "=", "value": true } ] ```. Example: [{"type": "string", "column": "name", "operator": "=", "value": "chat-completion"}].
- `from_start_time` (`string`, optional): Retrieve only observations with a start_time on or after this datetime (ISO 8601). Example: 2026-05-01T00:00:00Z.
- `is_root_observation` (`boolean`, optional): Filter by whether an observation is a logical root. Root observations include observations without a physical parent and observations marked as app roots by the SDK. An app-root observation may have `isRootObservation=true` and a non-null `parentObservationId`. Example: true.
- `level` (`string`, optional): Optional filter for observations with a specific level (e.g. "DEBUG", "DEFAULT", "WARNING", "ERROR"). Example: ERROR. One of: `DEBUG`, `DEFAULT`, `WARNING`, `ERROR`.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `name` (`string`, optional): Name. Example: chat-completion.
- `parent_observation_id` (`string`, optional): Filter by the physical parent observation ID. An empty value matches only observations without a physical parent. Use `isRootObservation` to include observations marked as app roots by the SDK, which may retain a non-null `parentObservationId`. Example: obs-1.
- `parse_io_as_json` (`boolean`, optional): **Deprecated.** Setting this to `true` will return a 400 error. Input/output fields are always returned as raw strings. Remove this parameter or set it to `false`. Example: false.
- `session_id` (`string`, optional): Filter by session ID. Example: session-abc.
- `to_start_time` (`string`, optional): Retrieve only observations with a start_time before this datetime (ISO 8601). Example: 2026-05-31T23:59:59Z.
- `trace_id` (`string`, optional): Trace ID. Example: 4bf92f3577b34da6a3ce929d0e0e4736.
- `type` (`string`, optional): Filter by observation type (e.g., "GENERATION", "SPAN", "EVENT", "AGENT", "TOOL", "CHAIN", "RETRIEVER", "EVALUATOR", "EMBEDDING", "GUARDRAIL"). Example: GENERATION.
- `user_id` (`string`, optional): User ID. Example: user-123.
- `version` (`string`, optional): Optional filter to only include observations with a certain version. Example: 1.0.0.

Example input: `{}`

### `langfuse_project_get`

Get Project · Read-only

Get the project the API keys belong to.
Returns the project's ID, name and settings.
Use the ID where tools ask for a project_id.

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

Inputs: none.

Example input: `{}`

### `langfuse_prompt_get`

Get Prompt · Read-only

Get a prompt by name, at a version or label (production by default).
Returns the prompt text or chat messages, config, version, labels and tags.

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

Inputs:

- `prompt_name` (`string`, required): The name of the prompt. If the prompt is in a folder (e.g., "folder/subfolder/prompt-name"), the folder path must be URL encoded. Example: support-answer.
- `label` (`string`, optional): Label of the prompt to be retrieved. Defaults to "production" if no label or version is set. Example: production.
- `resolve` (`boolean`, optional): Resolve prompt dependencies before returning the prompt. Defaults to `true`. Set to `false` to return the raw stored prompt with dependency tags intact. This bypasses prompt caching and is intended for debugging or one-off jobs, not production runtime fetches. Example: true.
- `version` (`integer`, optional): Version of the prompt to be retrieved. Example: 3.

Example input: `{"prompt_name":"folder/subfolder/prompt-name"}`

### `langfuse_prompts_list`

List Prompts · Read-only

List prompts with their versions, labels and tags, with filters and paging.
Returns each prompt's name, versions, labels, tags and last update.
Use langfuse_prompt_get for a prompt's content.

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

Inputs:

- `from_updated_at` (`string`, optional): Optional filter to only include prompt versions created/updated on or after a certain datetime (ISO 8601). Example: 2026-05-01T00:00:00Z.
- `label` (`string`, optional): Label. Example: production.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `name` (`string`, optional): Name. Example: support-answer.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.
- `tag` (`string`, optional): Tag. Example: support.
- `to_updated_at` (`string`, optional): Optional filter to only include prompt versions created/updated before a certain datetime (ISO 8601). Example: 2026-05-31T23:59:59Z.

Example input: `{}`

### `langfuse_score_config_get`

Get Score Config · Read-only

Get one score configuration.
Returns its name, data type, range or categories and description.

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

Inputs:

- `config_id` (`string`, required): The unique langfuse identifier of a score config. Example: cm4t5y6u7i8o9p0a1s2d3f.

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

### `langfuse_score_configs_list`

List Score Configs · Read-only

List score configurations, which define the name, data type and allowed values of scores.
Returns each config's ID, name, data type, range or categories.

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

Inputs:

- `limit` (`integer`, optional): Items per page. Example: 50.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.

Example input: `{}`

### `langfuse_scores_list`

List Scores · Read-only

List scores with filters by name, source, data type, trace, session, observation, queue, value range and time, with cursor paging.
Returns each score's ID, name, value, data type, source, comment and target.

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

Inputs:

- `author_user_id` (`string`, optional): Comma-separated list of author user IDs to filter by. Example: user-1.
- `config_id` (`string`, optional): Comma-separated list of score config IDs to filter by. Example: cm4t5y6u7i8o9p0a1s2d3f.
- `cursor` (`string`, optional): Cursor from the previous response's meta, for the next page. Example: eyJ0IjoxfQ.
- `data_type` (`string`, optional): Comma-separated list of data types to filter by (NUMERIC, BOOLEAN, CATEGORICAL, TEXT, CORRECTION). Case-insensitive — `numeric` and `NUMERIC` are equivalent. Must be a single value when used with value, valueMin, or valueMax; otherwise the request returns HTTP 400. Must be NUMERIC when used with valueMin or valueMax. Example: NUMERIC.
- `environment` (`string`, optional): Comma-separated list of environments to filter by. Example: production.
- `experiment_id` (`string`, optional): Comma-separated list of dataset run IDs (experiment IDs) to filter by. Mutually exclusive with traceId, sessionId, observationId. Example: exp-1.
- `fields` (`string`, optional): Comma-separated field groups to include in addition to the always-returned core fields. Allowed: details, subject, annotation — see the endpoint description for the fields each group adds. Unknown names return HTTP 400. Example: core.
- `from_timestamp` (`string`, optional): Inclusive lower bound on the score timestamp. Example: 2026-05-01T00:00:00Z.
- `id` (`string`, optional): Comma-separated list of score IDs to filter by (OR within, AND across filters). Example: cm5g6h7j8k9l0z1x2c3v4b.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `name` (`string`, optional): Comma-separated list of score names to filter by. Example: helpfulness.
- `observation_id` (`string`, optional): Comma-separated list of observation IDs to filter by. Requires traceId to be specified, because observation IDs are scoped to a trace. Mutually exclusive with sessionId, experimentId. Returns HTTP 400 when used without traceId. Example: obs-1.
- `queue_id` (`string`, optional): ID of the annotation queue. Get it from langfuse_annotation_queues_list. Example: cm1q2w3e4r5t6y7u8i9o0p.
- `session_id` (`string`, optional): Comma-separated list of session IDs to filter by. Mutually exclusive with traceId, observationId, experimentId. Example: session-abc.
- `source` (`string`, optional): Comma-separated list of score sources to filter by (e.g. API, ANNOTATION, EVAL). Case-insensitive — `api` and `API` are equivalent. Example: API.
- `to_timestamp` (`string`, optional): Exclusive upper bound on the score timestamp. Example: 2026-05-31T23:59:59Z.
- `trace_id` (`string`, optional): Comma-separated list of trace IDs to filter by. Mutually exclusive with sessionId, experimentId. May be combined with observationId to scope the observation lookup to a specific trace. Example: 4bf92f3577b34da6a3ce929d0e0e4736.
- `value` (`string`, optional): Comma-separated list of exact values to filter by. Requires a single dataType from NUMERIC, BOOLEAN, or CATEGORICAL; any other dataType, multiple dataTypes, or omitting dataType returns HTTP 400. For BOOLEAN, each value must be "true" or "false"; for NUMERIC, each value must be a finite number. Otherwise the request returns HTTP 400. Example: 1.
- `value_max` (`number`, optional): Inclusive upper bound on the numeric value. Requires dataType=NUMERIC as a single value; otherwise the request returns HTTP 400. Example: 1.
- `value_min` (`number`, optional): Inclusive lower bound on the numeric value. Requires dataType=NUMERIC as a single value; otherwise the request returns HTTP 400. Example: 0.

Example input: `{}`

### `langfuse_skill_get`

Get Skill · Read-only

Get a skill's metadata and file manifest at a version or label.
Returns the skill version and its files.

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

Inputs:

- `skill_name` (`string`, required): Skill Name. Example: summarize-ticket.
- `include_content` (`boolean`, optional): Include UTF-8 text content in each file. Defaults to false. Example: true.
- `label` (`string`, optional): Label. Example: example.
- `version` (`integer`, optional): Version. Example: 3.

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

### `langfuse_skills_list`

List Skills · Read-only

List skills with their latest version metadata.
Returns each skill's name, tags, labels and timestamps.

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

Inputs:

- `from_updated_at` (`string`, optional): Include skills whose latest version was updated at or after this timestamp. Example: 2026-01-01T00:00:00Z.
- `limit` (`integer`, optional): Items per page. Example: 50.
- `name` (`string`, optional): Filter by exact skill name. Example: support-answer.
- `page` (`integer`, optional): Page number, starting at 1. Example: 1.
- `search` (`string`, optional): Case-insensitive search across skill names and latest-version descriptions. Example: example.
- `tag` (`string`, optional): Filter by a shared skill tag. Example: example.
- `to_updated_at` (`string`, optional): Include skills whose latest version was updated before this timestamp. Example: 2026-01-01T00:00:00Z.

Example input: `{}`

### `langfuse_annotation_queue_assignment_create`

Assign User to Queue · Write

Assign a user to an annotation queue.
Returns the assignment.

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

Inputs:

- `queue_id` (`string`, required): ID of the annotation queue. Get it from langfuse_annotation_queues_list. Example: cm1q2w3e4r5t6y7u8i9o0p.
- `user_id` (`string`, required): Langfuse user ID to assign. Example: user-1.

Example input: `{"queue_id":"<queue_id>","user_id":"<user_id>"}`

### `langfuse_annotation_queue_create`

Create Annotation Queue · Write

Create an annotation queue for human review.
Returns the new queue.

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

Inputs:

- `name` (`string`, required): Name. Example: Support review.
- `score_config_ids` (`array`, required): Score Config IDs. Example: ["cm4t5y6u7i8o9p0a1s2d3f"].
- `description` (`string`, optional): Description. Example: Review flagged answers.

Example input: `{"name":"<name>","score_config_ids":["cm4t5y6u7i8o9p0a1s2d3f"]}`

### `langfuse_annotation_queue_item_add`

Add Annotation Queue Item · Write

Add a trace, observation or session to an annotation queue for review.
Returns the new queue item.

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

Inputs:

- `object_id` (`string`, required): Object ID. Example: 4bf92f3577b34da6a3ce929d0e0e4736.
- `object_type` (`string`, required): Object Type. Example: TRACE. One of: `TRACE`, `OBSERVATION`, `SESSION`.
- `queue_id` (`string`, required): ID of the annotation queue. Get it from langfuse_annotation_queues_list. Example: cm1q2w3e4r5t6y7u8i9o0p.
- `status` (`string`, optional): Defaults to PENDING for new queue items. Example: PENDING. One of: `PENDING`, `COMPLETED`.

Example input: `{"object_id":"<object_id>","object_type":"TRACE","queue_id":"<queue_id>"}`

### `langfuse_annotation_queue_item_update`

Update Annotation Queue Item · Write

Change the status of an annotation queue item, for example to COMPLETED.
Returns the updated item.

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

Inputs:

- `item_id` (`string`, required): ID of the queue item. Example: cm2a3s4d5f6g7h8j9k0l1z.
- `queue_id` (`string`, required): ID of the annotation queue. Get it from langfuse_annotation_queues_list. Example: cm1q2w3e4r5t6y7u8i9o0p.
- `status` (`string`, optional): Status. Example: COMPLETED. One of: `PENDING`, `COMPLETED`.

Example input: `{"item_id":"<item_id>","queue_id":"<queue_id>"}`

### `langfuse_comment_create`

Create Comment · Write

Comment on a trace, observation, session or prompt.
Returns the new comment's ID.

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

Inputs:

- `content` (`string`, required): The content of the comment. May include markdown. Currently limited to 5000 characters. Example: Hallucinated the refund policy.
- `object_id` (`string`, required): The id of the object to attach the comment to. If this does not reference a valid existing object, an error will be thrown. Example: 4bf92f3577b34da6a3ce929d0e0e4736.
- `object_type` (`string`, required): The type of the object to attach the comment to (trace, observation, session, prompt). Example: TRACE.
- `project_id` (`string`, required): ID of the project. Get it from langfuse_project_get. Example: cm9p0o1i2u3y4t5r6e7w8q.
- `author_user_id` (`string`, optional): The id of the user who created the comment. Must be a member of the organization that owns the project, otherwise an error will be thrown. Example: user-1.

Example input: `{"content":"<content>","object_id":"<object_id>","object_type":"<object_type>","project_id":"<project_id>"}`

### `langfuse_dataset_create`

Create Dataset · Write

Create a dataset for evaluations and experiments.
Returns the new dataset.

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

Inputs:

- `name` (`string`, required): Name. Example: golden-qa.
- `description` (`string`, optional): Description. Example: Golden questions for the support bot.
- `expected_output_schema` (`object`, optional): JSON Schema for validating dataset item expected outputs. When set, all new and existing dataset items will be validated against this schema. Example: {"type": "object"}.
- `input_schema` (`object`, optional): JSON Schema for validating dataset item inputs. When set, all new and existing dataset items will be validated against this schema. Example: {"type": "object"}.
- `metadata` (`object`, optional): Metadata. Example: {"owner": "ml-team"}.

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

### `langfuse_dataset_item_create`

Create Dataset Item · Write

Add an item to a dataset, or update it when the id already exists.
Returns the item.

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

Inputs:

- `dataset_name` (`string`, required): Dataset Name. Example: golden-qa.
- `expected_output` (`object`, optional): Expected output; any JSON value. Example: {"answer": "Use the Forgot password link."}.
- `id` (`string`, optional): Dataset items are upserted on their id. Id needs to be unique (project-level), cannot be reused across datasets, and must be at most 255 characters. Example: item-001.
- `input` (`object`, optional): Input of the item; any JSON value. Example: {"question": "How do I reset my password?"}.
- `metadata` (`object`, optional): Any JSON metadata. Example: {"source": "support"}.
- `source_observation_id` (`string`, optional): Source Observation ID. Example: obs-1.
- `source_trace_id` (`string`, optional): Source Trace ID. Example: 4bf92f3577b34da6a3ce929d0e0e4736.
- `status` (`string`, optional): Defaults to ACTIVE for newly created items. Example: ACTIVE. One of: `ACTIVE`, `ARCHIVED`.

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

### `langfuse_evaluation_rule_create`

Create Evaluation Rule · Write

Create an evaluation rule that runs evaluators on matching new data.
Returns the new rule.

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

Inputs:

- `enabled` (`boolean`, required): Whether live execution should start immediately. Enabled rules require at least one evaluator assignment. Example: true.
- `evaluator_assignments` (`array`, required): Evaluators to attach to this rule. Disabled rules may use an empty list as a draft. Example: [{"evaluatorId": "cm7o8p9a0s1d2f3g4h5j6k"}].
- `name` (`string`, required): Human-readable rule name. Names are not identifiers and do not need to be unique. Example: Helpfulness on prod.
- `filter` (`array`, optional): Conditions used to select observations. Defaults to an empty list, which matches every incoming observation. Example: [].
- `sampling` (`number`, optional): Fraction of matching observations to evaluate. Omit this field to use the default of `1`, which evaluates every match. `null` is not accepted. Example: 0.1.

Example input: `{"enabled":true,"evaluator_assignments":[{"evaluatorId":"cm7o8p9a0s1d2f3g4h5j6k"}],"name":"<name>"}`

### `langfuse_evaluation_rule_update`

Update Evaluation Rule · Write

Update an evaluation rule's name, enabled flag, sampling, filter or evaluators.
Returns the updated rule.

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

Inputs:

- `evaluation_rule_id` (`string`, required): Stable evaluation-rule identifier returned by the evaluation-rule endpoints. Example: cm6n7m8q9w0e1r2t3y4u5i.
- `enabled` (`boolean`, optional): New desired live-execution state. Example: false.
- `evaluator_assignments` (`array`, optional): Complete replacement assignment list. An empty list disables the rule. Example: [{"evaluatorId": "cm7o8p9a0s1d2f3g4h5j6k"}].
- `filter` (`array`, optional): Complete replacement filter list. An empty list matches every incoming observation. Example: [].
- `name` (`string`, optional): New human-readable rule name. Example: Helpfulness on prod.
- `sampling` (`number`, optional): New fraction of matching observations to evaluate. Omit to keep the current value. Example: 0.1.

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

### `langfuse_model_create`

Create Model · Write

Create a custom model definition so Langfuse can compute cost for its generations.
Returns the new model.

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

Inputs:

- `match_pattern` (`string`, required): Regex pattern which matches this model definition to generation.model. Useful in case of fine-tuned models. If you want to exact match, use `(?i)^modelname$`. Example: (?i)^(my-finetune)$.
- `model_name` (`string`, required): Name of the model definition. If multiple with the same name exist, they are applied in the following order: (1) custom over built-in, (2) newest according to startTime where model.startTime<observation.startTime. Example: my-finetune.
- `unit` (`string`, required): Unit used by this model. Example: TOKENS. One of: `CHARACTERS`, `TOKENS`, `MILLISECONDS`, `SECONDS`, `IMAGES`, `REQUESTS`.
- `input_price` (`number`, optional): Deprecated. Use 'pricingTiers' instead. Price (USD) per input unit. Creates a default tier if pricingTiers not provided. Example: 0.000002.
- `output_price` (`number`, optional): Deprecated. Use 'pricingTiers' instead. Price (USD) per output unit. Creates a default tier if pricingTiers not provided. Example: 0.000008.
- `start_date` (`string`, optional): Apply only to generations which are newer than this ISO date. Example: 2026-01-01T00:00:00Z.
- `tokenizer_config` (`object`, optional): Legacy tokenizer configuration retained for API compatibility. Ingestion no longer uses it to infer observation usage. Example: {"tokenizerModel": "gpt-4o"}.
- `tokenizer_id` (`string`, optional): Legacy tokenizer identifier retained for API compatibility. Ingestion no longer uses it to infer observation usage. Example: openai. One of: `openai`, `claude`.
- `total_price` (`number`, optional): Deprecated. Use 'pricingTiers' instead. Price (USD) per total units. Cannot be set if input or output price is set. Creates a default tier if pricingTiers not provided. Example: 0.00001.

Example input: `{"match_pattern":"<match_pattern>","model_name":"<model_name>","unit":"CHARACTERS"}`

### `langfuse_model_upsert`

Create or Replace Model · Write

Create or replace a project-owned model definition by ID.
Returns the model.
Built-in Langfuse models cannot be replaced.

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

Inputs:

- `match_pattern` (`string`, required): Regex pattern which matches this model definition to generation.model. Useful in case of fine-tuned models. If you want to exact match, use `(?i)^modelname$`. Example: (?i)^(my-finetune)$.
- `model_id` (`string`, required): ID of the model to create or replace. Example: cm0q1w2e3r4t5y6u7i8o9p.
- `model_name` (`string`, required): Name of the model definition. If multiple with the same name exist, they are applied in the following order: (1) custom over built-in, (2) newest according to startTime where model.startTime<observation.startTime. Example: my-finetune.
- `unit` (`string`, required): Unit used by this model. Example: TOKENS. One of: `CHARACTERS`, `TOKENS`, `MILLISECONDS`, `SECONDS`, `IMAGES`, `REQUESTS`.
- `input_price` (`number`, optional): Deprecated. Use 'pricingTiers' instead. Price (USD) per input unit. Creates a default tier if pricingTiers not provided. Example: 0.000002.
- `output_price` (`number`, optional): Deprecated. Use 'pricingTiers' instead. Price (USD) per output unit. Creates a default tier if pricingTiers not provided. Example: 0.000008.
- `start_date` (`string`, optional): Apply only to generations which are newer than this ISO date. Example: 2026-01-01T00:00:00Z.
- `tokenizer_config` (`object`, optional): Legacy tokenizer configuration retained for API compatibility. Ingestion no longer uses it to infer observation usage. Example: {"tokenizerModel": "gpt-4o"}.
- `tokenizer_id` (`string`, optional): Legacy tokenizer identifier retained for API compatibility. Ingestion no longer uses it to infer observation usage. Example: openai. One of: `openai`, `claude`.
- `total_price` (`number`, optional): Deprecated. Use 'pricingTiers' instead. Price (USD) per total units. Cannot be set if input or output price is set. Creates a default tier if pricingTiers not provided. Example: 0.00001.

Example input: `{"match_pattern":"<match_pattern>","model_id":"<model_id>","model_name":"<model_name>","unit":"CHARACTERS"}`

### `langfuse_prompt_create`

Create Prompt Version · Write

Create a new version of a text or chat prompt; a new name creates the prompt.
Returns the created prompt version.
Set labels to ["production"] to make it the version clients fetch.

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

Inputs:

- `name` (`string`, required): The full prompt name, including any folder path, with at most 255 characters. Example: support-answer.
- `prompt` (`string`, required): The prompt: a string for a text prompt, or a list of chat messages with role and content for a chat prompt. Example: "Answer the question: {{question}}".
- `commit_message` (`string`, optional): Commit message for this prompt version. Example: Shorter answer.
- `config` (`object`, optional): Any JSON config stored with the prompt, such as model and temperature. Example: {"model": "gpt-4o", "temperature": 0.2}.
- `labels` (`array`, optional): List of deployment labels of this prompt version. Example: ["staging"].
- `tags` (`array`, optional): List of tags to apply to all versions of this prompt. Example: ["support"].
- `type` (`string`, optional): text or chat; must match the shape of prompt. Example: text. One of: `text`, `chat`.

Example input: `{"name":"<name>","prompt":"Answer the question: {{question}}"}`

### `langfuse_prompt_version_labels_update`

Update Prompt Version Labels · Write

Add or move labels, such as production, to a prompt version.
Returns the updated prompt version.

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

Inputs:

- `name` (`string`, required): Name of the prompt. Example: support-answer.
- `version` (`integer`, required): Version number. Example: 3.
- `new_labels` (`array`, optional): New labels for the prompt version. Labels are unique across versions. The "latest" label is reserved and managed by Langfuse. Example: ["production"].
- `remove_labels` (`array`, optional): Labels to remove from the prompt version. The "latest" label is reserved and managed by Langfuse. Example: ["staging"].

Example input: `{"name":"<name>","version":3}`

### `langfuse_score_config_create`

Create Score Config · Write

Create a score configuration for consistent scoring.
Returns the new config.

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

Inputs:

- `data_type` (`string`, required): Data Type. Example: NUMERIC. One of: `NUMERIC`, `BOOLEAN`, `CATEGORICAL`, `TEXT`.
- `name` (`string`, required): Name of the score config. Max 80 characters. Only letters, numbers, underscores, spaces, periods, parentheses, and hyphens are allowed. Example: helpfulness.
- `categories` (`array`, optional): Configure custom categories for categorical scores. Pass a list of objects with `label` and `value` properties. Categories are autogenerated for boolean configs and cannot be passed. Example: [{"label": "good", "value": 1}, {"label": "bad", "value": 0}].
- `description` (`string`, optional): Description is shown across the Langfuse UI and can be used to e.g. explain the config categories in detail, why a numeric range was set, or provide additional context on config name or usage. Example: How helpful the answer was.
- `max_value` (`number`, optional): Configure a maximum value for numerical scores. If not set, the maximum value defaults to +∞. Example: 1.
- `min_value` (`number`, optional): Configure a minimum value for numerical scores. If not set, the minimum value defaults to -∞. Example: 0.

Example input: `{"data_type":"NUMERIC","name":"<name>"}`

### `langfuse_score_config_update`

Update Score Config · Write

Update a score configuration's description, range, categories or archive state.
Returns the updated config.

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

Inputs:

- `config_id` (`string`, required): The unique langfuse identifier of a score config. Example: cm4t5y6u7i8o9p0a1s2d3f.
- `categories` (`array`, optional): Configure custom categories for categorical scores. Pass a list of objects with `label` and `value` properties. Categories are autogenerated for boolean configs and cannot be passed. Example: [{"label": "good", "value": 1}, {"label": "bad", "value": 0}].
- `description` (`string`, optional): Description is shown across the Langfuse UI and can be used to e.g. explain the config categories in detail, why a numeric range was set, or provide additional context on config name or usage. Example: How helpful the answer was.
- `is_archived` (`boolean`, optional): The status of the score config showing if it is archived or not. Example: false.
- `max_value` (`number`, optional): Configure a maximum value for numerical scores. If not set, the maximum value defaults to +∞. Example: 1.
- `min_value` (`number`, optional): Configure a minimum value for numerical scores. If not set, the minimum value defaults to -∞. Example: 0.
- `name` (`string`, optional): Name of the score config. Max 80 characters. Only letters, numbers, underscores, spaces, periods, parentheses, and hyphens are allowed. Example: support-answer.

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

### `langfuse_scores_create`

Create Scores · Write

Create one or more scores on traces, observations, sessions or dataset runs, such as user feedback or evaluation results.
Returns the IDs of the accepted scores; scores are written asynchronously.

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

Inputs:

- `scores` (`array`, required): Scores to create. Each needs a name and a value (a number, a category string or a boolean) and a target: traceId (optionally with observationId), sessionId or datasetRunId. Optional: dataType (NUMERIC, CATEGORICAL, BOOLEAN), comment, configId to validate against a score config, environment, metadata and id (to update an existing score). Example: [{"name": "helpfulness", "value": 0.9, "traceId": "4bf92f3577b34da6a3ce929d0e0e4736", "comment": "Good answer"}].

Example input: `{"scores":[{"name":"helpfulness","value":0.9,"traceId":"4bf92f3577b34da6a3ce929d0e0e4736","comment":"Good answer"}]}`

### `langfuse_annotation_queue_item_delete`

Remove Annotation Queue Item · Destructive

Remove an item from an annotation queue.
Returns a confirmation.

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

Inputs:

- `item_id` (`string`, required): ID of the queue item. Example: cm2a3s4d5f6g7h8j9k0l1z.
- `queue_id` (`string`, required): ID of the annotation queue. Get it from langfuse_annotation_queues_list. Example: cm1q2w3e4r5t6y7u8i9o0p.

Example input: `{"item_id":"<item_id>","queue_id":"<queue_id>"}`

### `langfuse_dataset_item_delete`

Delete Dataset Item · Destructive

Delete a dataset item and its run items.
Returns a confirmation message.

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

Inputs:

- `dataset_item_id` (`string`, required): ID of the dataset item. Example: item-001.

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

### `langfuse_evaluation_rule_delete`

Delete Evaluation Rule · Destructive

Delete an evaluation rule. Past results stay.
Returns an empty response on success.

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

Inputs:

- `evaluation_rule_id` (`string`, required): Stable evaluation-rule identifier returned by the evaluation-rule endpoints. Example: cm6n7m8q9w0e1r2t3y4u5i.

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

### `langfuse_evaluator_delete`

Delete Evaluator · Destructive

Delete an evaluator and all its versions.
Returns an empty response on success.

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

Inputs:

- `evaluator_id` (`string`, required): Stable evaluator identifier returned by the evaluator endpoints. Example: cm7o8p9a0s1d2f3g4h5j6k.

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

### `langfuse_llm_connection_delete`

Delete LLM Connection · Destructive

Delete a project LLM connection. Evaluators that use it stop working.
Returns an empty response on success.

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

Inputs:

- `llm_connection_id` (`string`, required): ID of the LLM connection. Get it from langfuse_llm_connections_list. Example: cm1z2x3c4v5b6n7m8a9s0d.

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

### `langfuse_model_delete`

Delete Model · Destructive

Delete a project-owned model definition.
Returns an empty response on success.

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

Inputs:

- `model_id` (`string`, required): ID of the model. Get it from langfuse_models_list. Example: cm0q1w2e3r4t5y6u7i8o9p.

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

### `langfuse_prompt_delete`

Delete Prompt Versions · Destructive

Delete a prompt, or only the versions with a given version number or label.
Returns an empty response on success.

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

Inputs:

- `prompt_name` (`string`, required): The name of the prompt. Example: support-answer.
- `label` (`string`, optional): Optional label to filter deletion. If specified, deletes all prompt versions that have this label. Example: staging.
- `version` (`integer`, optional): Optional version to filter deletion. If specified, deletes only this specific version of the prompt. Example: 3.

Example input: `{"prompt_name":"folder/subfolder/prompt-name"}`

### `langfuse_score_delete`

Delete Score · Destructive

Delete a score.
Returns an empty response on success.

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

Inputs:

- `score_id` (`string`, required): The unique langfuse identifier of a score. Example: cm5g6h7j8k9l0z1x2c3v4b.

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

### `langfuse_trace_delete`

Delete Trace · Destructive

Delete one trace and its observations.
Returns a confirmation message.
Use langfuse_traces_delete for several traces.

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

Inputs:

- `trace_id` (`string`, required): ID of the trace to delete. Example: 4bf92f3577b34da6a3ce929d0e0e4736.

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

### `langfuse_traces_delete`

Delete Traces · Destructive

Delete several traces and their observations.
Returns a confirmation message.

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

Inputs:

- `trace_ids` (`array`, required): IDs of the traces to delete. Example: ["4bf92f3577b34da6a3ce929d0e0e4736", "5ce03a4688c45eb7b4df030e1f1f5847"].

Example input: `{"trace_ids":["4bf92f3577b34da6a3ce929d0e0e4736","5ce03a4688c45eb7b4df030e1f1f5847"]}`

## 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 |
