Athina AI

Athina AI is an LLM monitoring, evaluation, and experimentation platform for building production-grade AI applications. Its REST API lets teams log inferences and traces, manage datasets, run 50+ preset and custom evaluations, version and run prompt templates, and collaborate on experiments across the full LLM development lifecycle.

5 APIs 0 Features

AILLMObservabilityEvaluationMonitoring

APIs

Athina Logging & Inferences API

Log LLM inferences and prompt runs (prompt, response, model, tokens, cost, context, customer/session metadata), update logs by id or external reference, and build nested traces ...

Athina Datasets API

Create datasets, list and retrieve datasets, add up to 1000 rows (query, context, response, expected_response and custom fields), update individual cells, and delete datasets us...

Athina Evaluations API

Run single evaluations or suites of 50+ preset and custom evals (RAG, summarization, function-calling, safety) against datasets and logged inferences, with results surfaced in t...

Athina Prompts API

Create and version prompt templates by slug, fetch the default version, and run a prompt by slug with input variables, model, and parameters (temperature, max_tokens, top_p, pen...

Athina Experiments API

Compare prompts, models, and parameters across datasets by combining versioned prompt runs with evaluations, enabling side-by-side experimentation over the prompt and dataset APIs.

Collections

Athina AI API

OPEN

Pricing Plans

Athina Plans Pricing

3 plans

PLANS

Rate Limits

Athina Rate Limits

4 limits

RATE LIMITS

FinOps

Athina Finops

FINOPS

Resources

👥

GitHubOrganization

Sources

opencollection: 1.0.0
info:
  name: Athina AI API
  version: '1.0'
request:
  auth:
    type: apikey
    apikey:
      key: athina-api-key
      value: '{{athinaApiKey}}'
      in: header
items:
- info:
    name: Logging & Inferences
    type: folder
  items:
  - info:
      name: Log an LLM inference
      type: http
    http:
      method: POST
      url: https://log.athina.ai/api/v1/log/inference
      body:
        type: json
        data: '{"language_model_id":"gpt-4o","prompt":[{"role":"user","content":"What is the capital of France?"}],"response":"Paris"}'
    docs: Log a single LLM inference / prompt run to Athina.
  - info:
      name: Update a log by id
      type: http
    http:
      method: PUT
      url: https://api.athina.ai/api/v1/logs/{log_id}
      body:
        type: json
        data: '{"expected_response":"Paris"}'
    docs: Update a logged inference by its Athina log id.
  - info:
      name: Update a log by external reference id
      type: http
    http:
      method: PUT
      url: https://api.athina.ai/api/v1/logs/external/{external_reference_id}
      body:
        type: json
        data: '{"expected_response":"Paris"}'
    docs: Update a logged inference by external_reference_id.
- info:
    name: Tracing
    type: folder
  items:
  - info:
      name: Create a trace
      type: http
    http:
      method: POST
      url: https://api.athina.ai/api/v1/trace/
      body:
        type: json
        data: '{"name":"my-app-run","start_time":"2026-06-20T12:00:00Z","status":"ok","attributes":{}}'
    docs: Create a new trace for an application run.
  - info:
      name: Get a trace
      type: http
    http:
      method: GET
      url: https://api.athina.ai/api/v1/trace/{trace_id}
    docs: Retrieve a trace and its spans by id.
  - info:
      name: Update a trace
      type: http
    http:
      method: PUT
      url: https://api.athina.ai/api/v1/trace/{trace_id}
      body:
        type: json
        data: '{"name":"my-app-run","start_time":"2026-06-20T12:00:00Z","end_time":"2026-06-20T12:00:05Z"}'
    docs: Update a trace name and timing.
  - info:
      name: Create a span
      type: http
    http:
      method: POST
      url: https://api.athina.ai/api/v1/trace/{trace_id}/spans
      body:
        type: json
        data: '{"name":"llm-call","trace_id":"{trace_id}","span_type":"generation","start_time":"2026-06-20T12:00:01Z","status":"ok","attributes":{}}'
    docs: Create a span within a trace (use span_type generation for LLM calls).
  - info:
      name: Get a span
      type: http
    http:
      method: GET
      url: https://api.athina.ai/api/v1/trace/{trace_id}/spans/{span_id}
    docs: Retrieve a single span within a trace.
  - info:
      name: Update a span
      type: http
    http:
      method: PUT
      url: https://api.athina.ai/api/v1/trace/{trace_id}/spans/{span_id}
      body:
        type: json
        data: '{"name":"llm-call","span_type":"generation","start_time":"2026-06-20T12:00:01Z","end_time":"2026-06-20T12:00:03Z"}'
    docs: Update a span name, type, and timing.
- info:
    name: Datasets
    type: folder
  items:
  - info:
      name: Create a dataset
      type: http
    http:
      method: POST
      url: https://api.athina.ai/api/v1/dataset_v2
      body:
        type: json
        data: '{"source":"api","name":"My Dataset"}'
    docs: Create a new dataset, optionally seeding prompt template and rows.
  - info:
      name: Get a dataset
      type: http
    http:
      method: GET
      url: https://api.athina.ai/api/v1/dataset_v2/{dataset_id}
    docs: Retrieve a dataset and its rows by id.
  - info:
      name: Delete a dataset
      type: http
    http:
      method: DELETE
      url: https://api.athina.ai/api/v1/dataset_v2/{dataset_id}
    docs: Delete a dataset by id.
  - info:
      name: Add rows to a dataset
      type: http
    http:
      method: POST
      url: https://api.athina.ai/api/v1/dataset_v2/{dataset_id}/add-rows
      body:
        type: json
        data: '{"dataset_rows":[{"query":"What is the capital of Spain?","response":"Madrid","expected_response":"Madrid"}]}'
    docs: Append up to 1000 rows to an existing dataset.
  - info:
      name: Update a cell in a dataset
      type: http
    http:
      method: POST
      url: https://api.athina.ai/api/v1/dataset_v2/{dataset_id}/cell
      body:
        type: json
        data: '{"row_no":0,"column_name":"expected_response","value":"Paris"}'
    docs: Update the value of a single dataset cell.
- info:
    name: Evaluations
    type: folder
  items:
  - info:
      name: Run an evaluation
      type: http
    http:
      method: POST
      url: https://api.athina.ai/api/v1/eval/run
      body:
        type: json
        data: '{"dataset_id":"","evals":[{"name":"DoesResponseAnswerQuery","type":"llm_grader"}],"model":"gpt-4o"}'
    docs: Run one or more preset or custom evals against a dataset or rows.
- info:
    name: Prompts
    type: folder
  items:
  - info:
      name: Create or version a prompt template
      type: http
    http:
      method: PUT
      url: https://api.athina.ai/api/v1/prompt/{prompt_slug}
      body:
        type: json
        data: '{"prompt":[{"role":"user","content":"Summarize: {{text}}"}],"model":"gpt-4o","parameters":{"temperature":0.2},"commit_message":"Initial
          version"}'
    docs: Create a prompt template or a new incremented version at the given slug.
  - info:
      name: Get the default prompt version
      type: http
    http:
      method: GET
      url: https://api.athina.ai/api/v1/prompt/{prompt_slug}/default
    docs: Fetch the default version of a prompt template by slug.
  - info:
      name: Run a prompt
      type: http
    http:
      method: POST
      url: https://api.athina.ai/api/v1/prompt/{prompt_slug}/run
      body:
        type: json
        data: '{"variables":{"text":"Hello world"},"model":"gpt-4o","parameters":{"temperature":0.2,"max_tokens":256}}'
    docs: Execute a managed prompt template by slug with input variables.