Build web-aware AI apps with SRCX.
SRCX gives developers a fast API to search the web, extract clean content, and generate structured context for LLMs and agents.
Quickstart
1. Get a key
New accounts receive 3,000 developer credits for real testing.
2. Search
Use Search for ranked web sources with freshness and quality signals.
3. Give agents context
Use Context or MCP tools to pass cited source blocks into your AI workflow.
Authentication
Send your API key as a Bearer token or X-API-Key.
Authorization: Bearer lmsk_YOUR_API_KEY
Search API
Use Search for source discovery. SRCX ranking adapts to entities, freshness, tutorials, comparisons, and multilingual queries.
| Parameter | Description |
|---|---|
q | Query text, 1-500 chars. |
limit | 1-20 results. |
language | Optional language filter. |
freshness | day, week, month, year. |
Extract API
Extract one URL into markdown and plain text. Every URL passes SSRF and robots checks.
Context API
Context returns deduplicated, cited source blocks designed for models, agents, and RAG workflows.
MCP / Agent Tools
For Claude Desktop, Cursor-style agents, local models, and workflow builders, use the SRCX MCP server template. It exposes srcx_search, srcx_extract, and srcx_context as callable tools.
Credits
| Action | Credits |
|---|---|
/v1/search | 1 |
/v1/extract | 2 |
/v1/context | 5 |
New accounts receive 3,000 one-time developer credits. Paid plans refresh monthly.
Errors
All API responses use the same JSON envelope.
{
"ok": false,
"data": null,
"error": {
"code": "unauthorized",
"message": "Invalid API key",
"details": null
},
"request_id": "..."
}