API reference

Connect LLM Pulse to your stack

The REST API exposes AI visibility metrics, prompt responses, citations, recommendations, Search Console data, webhooks, and reporting data. The MCP endpoint exposes the same product surface to AI clients through OAuth or API-key authentication.

OpenAPI JSON Swagger JSON Live docs

Base URL

https://api.llmpulse.ai/api/v1

All REST examples below assume this base path.

Authentication

Authorization: Bearer llmpulse_your_api_key_here

REST API keys are available on Scale and Enterprise. MCP via OAuth is available on every plan, including Starter and free trial accounts.

Quickstart

Use the projects endpoint first, then request metrics for a project ID.

1. Validate your key

curl https://api.llmpulse.ai/api/v1/ping \
  -H "Authorization: Bearer llmpulse_your_api_key_here"

2. List projects

curl https://api.llmpulse.ai/api/v1/dimensions/projects \
  -H "Authorization: Bearer llmpulse_your_api_key_here"

3. Get summary metrics

curl "https://api.llmpulse.ai/api/v1/metrics/summary?project_id=123&metrics=mentions,visibility,weighted_visibility&range=30" \
  -H "Authorization: Bearer llmpulse_your_api_key_here"

4. Compare share of voice

curl "https://api.llmpulse.ai/api/v1/metrics/sov?project_id=123&range=30" \
  -H "Authorization: Bearer llmpulse_your_api_key_here"

Endpoint Groups

Representative endpoint map from the current OpenAPI spec. Use the linked JSON files for exact request and response schemas.

GET /metrics/timeseries

Daily, weekly, or monthly series for mentions, citations, responses, mention rate, citation rate, AI visibility score, positions, and sentiment.

GET /metrics/summary

Aggregated dashboard totals, min, max, last values, and position distribution for a selected date range.

GET /metrics/sov

Share of Voice comparison across your project and tracked competitors.

GET /metrics/top_sources

Domains and URLs most frequently cited in AI responses.

GET /dimensions/projects

Projects available to the current API key. Most analytics calls start with one of these project IDs.

GET /dimensions/prompts

Tracked prompts, tags, locale filters, models, and prompt metadata.

GET /answers

AI responses with search support. Retrieve full responses with mentions, citations, sources, entities, and related metadata by ID.

GET /citation_intelligence/groups

Group citations by URL, domain, or host with model breakdowns, citation rate, and average citation position.

GET /recommendations

Read recommendation runs and items that help teams prioritize content and visibility work.

POST /prompts

Bulk-create prompts, up to 100 per request. Requires a read-write API key.

POST /webhooks

Create signed outbound webhooks for project events. Available on Scale and Enterprise.

POST /mcp

JSON-RPC MCP endpoint for AI clients. OAuth is the recommended connection flow for interactive AI assistants.

Common Parameters

Parameter Use
project_id Required for most project-scoped endpoints.
range Shortcut date range, commonly 7, 30, 90, or 365 days.
from and to ISO8601 date range when you need exact boundaries.
model Filter by model, such as ChatGPT, Perplexity, Gemini, Google AI Overview, or Google AI Mode.
country_code and language_code Filter localized prompt runs and analytics.
page and per_page Paginate list endpoints. Most list endpoints cap page size at 100.

MCP in One Minute

Interactive assistant use: paste https://api.llmpulse.ai/api/v1/mcp into a compatible AI client and complete OAuth.

Headless use: send an API key in the bearer header. This path requires Scale or Enterprise.

Tool availability follows the user's plan and permissions, so locked product areas are filtered from the tool list.

Error Format

All API errors return a consistent envelope with a stable code and request ID.

{
  "error": {
    "code": "ERR_INVALID_API_KEY",
    "message": "API key is invalid or not found",
    "meta": {}
  },
  "request_id": "req_..."
}