Ian's Intel API

Programmatic, read-only access to the same curated crypto intelligence that publishes to the Ian's Intel channel — filtered news with original-source attribution, weekly briefs, and the raw news firehose.

• Base URL: https://api.iansintel.com/v1 (also reachable at https://iansintel.com/v1)
• Format: JSON over HTTPS. GET-only (read-only).
• Auth: API key via Authorization: Bearer header.
• Machine spec: /v1/openapi.json (OpenAPI 3.1 — point your tooling here).
• Freshness: curated posts and the firehose are real-time (poll every few seconds at most); weekly briefs post once a week.

Free while in preview. The API is currently free and full-featured. Please use a sensible request rate; heavy or commercial use may move to a paid tier in future — we'll give notice.

Quickstart

curl -H "Authorization: Bearer iik_live_YOUR_KEY" \
  "https://api.iansintel.com/v1/posts?limit=2"

{
  "data": [
    {
      "uid": "cp_1291",
      "id": 1291,
      "source": "Telegram",
      "username": "iansintel",
      "body": "COINBASE'S BASE NETWORK MOVES AWAY FROM OPTIMISM STACK TO CREATE THEIR OWN",
      "url": "https://t.me/iansintel/7812",
      "original_source_url": "https://base.org/blog/...",
      "created_at": "2026-02-18T17:14:02.105Z"
    }
  ],
  "pagination": { "next_cursor": "1290", "has_more": true },
  "meta": { "api_version": "v1", "request_id": "req_…", "tier": "free" }
}

Authentication

Every endpoint except /health and /openapi.json requires an API key:

Authorization: Bearer iik_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Keys are issued by Ian's Intel — contact us to request one. Keep your key secret; treat it like a password. If a key is exposed, ask us to revoke and reissue it.

Missing or invalid keys return 401:

{ "error": { "code": "unauthorized", "message": "Invalid or revoked API key." }, "meta": { … } }

Rate limits

• ~120 requests/minute per key. Exceeding it returns 429 with a Retry-After: 60 header.
• Watch the RateLimit-Remaining response header to pace yourself.
• The API is a live-updating feed — poll politely (every few seconds at most). For most uses, paginating on new data with since/cursor is far more efficient than re-fetching.

Response envelope

Every response shares one shape:

Errors replace data with error:

{ "error": { "code": "not_found", "message": "Post not found." }, "meta": { … } }

Pagination

List endpoints use keyset (cursor) pagination, newest first:

1. Make a request (optionally with limit, default 50, max 100).
2. If pagination.has_more is true, pass pagination.next_cursor back as the cursor parameter to get the next (older) page.
3. Stop when has_more is false.

# page 1
curl -H "Authorization: Bearer $KEY" "https://api.iansintel.com/v1/posts?limit=50"
# page 2 — use next_cursor from page 1
curl -H "Authorization: Bearer $KEY" "https://api.iansintel.com/v1/posts?limit=50&cursor=1290"

Rows are never skipped or duplicated across pages. Every item carries a stable uid (e.g. cp_1291) — use it to de-duplicate or reference items across sessions. To pull only new items since your last sync, keep the highest id/uid you've seen and pass since (a timestamp) or walk cursor until you reach it.

Endpoints

GET /v1/posts — curated feed

The filtered, aggregated news feed (with original-source links). This is the flagship dataset.

Post fields: uid, id, source, username, body, url (Telegram permalink), original_source_url (original article/tweet), created_at.

GET /v1/posts/{id} — single post

Returns one post by numeric id. include_raw supported. 404 if not found.

GET /v1/briefs — weekly briefs (list)

Metadata for the weekly intelligence briefs (no body). Paginated. Brief fields: uid (e.g. wb_12), id, week_start, week_end, post_count, created_at.

GET /v1/briefs/{id}

The full brief, including content (Markdown). 404 if not found.

GET /v1/firehose — raw news stream

The unfiltered upstream news firehose. Higher volume and less curated than /v1/posts.

Firehose fields: uid (e.g. tm_5678), id, pane, source, source_name, name, username, title, body, summary_ai (if requested), important, url, original_source_url, created_at.

GET /v1/sources — discovery

Returns { curated_sources, firehose_panes } so you can see the valid filter values.

GET /v1/health · GET /v1/openapi.json — open (no key)

Health/version, and the machine-readable OpenAPI spec.

Using it with an LLM / agent

The API is designed for AI agents. Point any OpenAPI-aware tool at https://api.iansintel.com/v1/openapi.json, and use the stable uids to reference items across turns. Keep responses token-light with include_body=false / include_raw=false (the default) until you need the full text.

MCP server

Plug Ian's Intel into Claude Desktop, Cursor, or any MCP client with our single-file MCP server — a thin, read-only bridge to this API. No install, no dependencies; just Node 18+.

1. Download it (iansintel-mcp.mjs):

curl -o iansintel-mcp.mjs https://iansintel.com/iansintel-mcp.mjs

2. Point your MCP client at it (use the path where you saved it):

{
  "mcpServers": {
    "iansintel": {
      "command": "node",
      "args": ["/absolute/path/to/iansintel-mcp.mjs"],
      "env": { "IANSINTEL_API_KEY": "iik_live_YOUR_KEY" }
    }
  }
}

It exposes tools (get_posts, get_post, get_briefs, get_brief, get_firehose, get_sources) and prompts (whats_moving, weekly_brief).

Stability

v1 is stable. We may add fields or endpoints without notice; we won't remove or change existing ones without shipping a new version (v2). Check meta.api_version.

Questions or a key request: t.me/cryptoian.