API Docs — AgentRank

Your agent's recommendations are only as good as its data. Don't let it guess from stale training data — give it the live index.

Query the AgentRank index programmatically. Free endpoints are unauthenticated and CORS-enabled. Pro endpoints require an API key for higher rate limits and additional features.

Base URL: https://agentrank-ai.com

Use AgentRank from your AI agent

The AgentRank MCP server gives Claude and other AI agents direct access to the live ranked index. No API key needed.

Install

claude mcp add agentrank -- npx -y agentrank-mcp-server

Or run directly

npx agentrank-mcp-server

The MCP server wraps the same endpoints documented below. Once installed, your agent can search, lookup, and get rankings without writing any API calls.

IDE & Editor Configs

Drop a config file into your project and your AI coding assistant will automatically query AgentRank for live tool recommendations instead of relying on stale training data.

Cursor .cursorrules

Copy to your project root or ~/.cursorrules for global use.

curl -o .cursorrules https://raw.githubusercontent.com/superlowburn/agentrank/main/skill/agentrank/cursor/.cursorrules

Preview

Instructs Cursor to query agentrank-ai.com/api/v1/search whenever you ask for tool or MCP server recommendations. Includes trigger conditions and result formatting guidance.

VS Code + Copilot .github/copilot-instructions.md

Copy to .github/copilot-instructions.md in your repo root. Copilot reads this automatically.

mkdir -p .github && curl -o .github/copilot-instructions.md https://raw.githubusercontent.com/superlowburn/agentrank/main/skill/agentrank/vscode/copilot-instructions.md

Preview

Instructs GitHub Copilot to query the AgentRank API for live tool data when recommending MCP servers, agent tools, or AI skills.

Cline .clinerules

Copy to your project root or ~/.clinerules for global use.

curl -o .clinerules https://raw.githubusercontent.com/superlowburn/agentrank/main/skill/agentrank/cline/.clinerules

Preview

Instructs Cline to call the AgentRank API (via curl or browser_action) for tool recommendations. Covers trigger conditions, query format, and result presentation.

All configs are in the agentrank GitHub repo. Copy, fork, or adapt them for your setup.

Pro API

Pro endpoints unlock higher rate limits, rank history, batch lookups, and richer signal data. All Pro endpoints require a valid API key.

Getting an API Key

API keys are available on the Pro and Enterprise plans. See pricing for tier details and sign-up. Keys come in three tiers:

Tier	Daily Limit	Batch	Rank History
`free`	100 req/day	No	No
`pro`	10,000 req/day	Yes	Yes
`enterprise`	Unlimited	Yes	Yes

Authentication

Pass your key in the Authorization header:

Authorization: Bearer ark_pro_a1b2c3d4e5f6...

Rate Limit Headers

All authenticated responses include rate limit headers (omitted for enterprise):

X-RateLimit-Limit: 10000
X-RateLimit-Remaining: 9987
X-RateLimit-Reset: 2026-03-17T23:59:59.999Z
X-API-Key-Tier: pro

Rate limits are per-key daily (midnight UTC reset). Free tier keys return 429 when exceeded.

V1 Endpoints

Full-featured endpoints with rate limiting (100 req/min per IP), rich signal data, and pagination.

GET /api/v1/search

Search tools, skills, and agents. Supports filtering by category, sorting, and pagination.

Parameters

Param	Type	Required	Description
`q`	string	No	Search query. Omit for full index.
`category`	string	No	`tool`, `skill`, or `agent`. Omit for all.
`sort`	string	No	`score` (default), `rank`, `stars`, `freshness`
`limit`	number	No	1–100, default 20
`offset`	number	No	Pagination offset, default 0

Example

curl "https://agentrank-ai.com/api/v1/search?q=playwright&category=tool&limit=3"

Response

{
  "query": "playwright",
  "category": "tool",
  "sort": "score",
  "results": [
    {
      "type": "tool",
      "id": "microsoft--playwright-mcp",
      "name": "microsoft/playwright-mcp",
      "description": "Playwright Tools for MCP",
      "score": 78.1,
      "rank": 34,
      "url": "https://agentrank-ai.com/tool/microsoft--playwright-mcp/",
      "raw": {
        "stars": 4200,
        "lastCommitAt": "2026-03-10T00:00:00.000Z",
        "language": "TypeScript",
        "license": "Apache-2.0"
      }
    }
  ],
  "meta": { "total": 1, "limit": 3, "offset": 0 }
}

Cache: 1min CDN / 30s browser • Rate limit: 100 req/min

GET /api/v1/tool/{id}

Detailed signal breakdown for a single tool or skill. Use owner--repo format for GitHub tools (replace / with --).

Examples

curl "https://agentrank-ai.com/api/v1/tool/microsoft--playwright-mcp"
curl "https://agentrank-ai.com/api/v1/tool/glama-playwright-mcp-server"  # skill by slug

Response

{
  "type": "tool",
  "id": "microsoft--playwright-mcp",
  "name": "microsoft/playwright-mcp",
  "description": "Playwright Tools for MCP",
  "score": 78.1,
  "rank": 34,
  "githubUrl": "https://github.com/microsoft/playwright-mcp",
  "signals": {
    "stars": 0.862,
    "freshness": 0.943,
    "issueHealth": 0.712,
    "contributors": 0.634,
    "dependents": 0.218,
    "descriptionQuality": 0.7,
    "licenseHealth": 1.0
  },
  "weights": {
    "stars": 0.12, "freshness": 0.23, "issueHealth": 0.24,
    "contributors": 0.09, "dependents": 0.24,
    "descriptionQuality": 0.05, "licenseHealth": 0.03
  },
  "raw": {
    "stars": 4200, "forks": 312, "contributors": 28,
    "openIssues": 14, "closedIssues": 89, "dependents": 7,
    "lastCommitAt": "2026-03-10T00:00:00.000Z",
    "language": "TypeScript", "license": "Apache-2.0",
    "isArchived": false, "topics": ["mcp", "playwright", "browser-automation"]
  }
}

Cache: 5min CDN / 1min browser • Rate limit: 100 req/min

GET /api/v1/movers

Tools and skills with the largest rank changes over a given period. Good for "biggest movers" content.

Parameters

Param	Type	Required	Description
`days`	number	No	Lookback window in days (1–30, default 7)
`limit`	number	No	1–50, default 10
`type`	string	No	`tool` (default), `skill`, or omit for both

Example

curl "https://agentrank-ai.com/api/v1/movers?days=7&limit=5&type=tool"

Response

{
  "movers": [
    {
      "full_name": "microsoft/playwright-mcp",
      "tool_type": "tool",
      "current_rank": 34,
      "prev_rank": 51,
      "rank_delta": 17,
      "current_score": 78.1,
      "stars": 4200,
      "url": "https://agentrank-ai.com/tool/microsoft--playwright-mcp/"
    }
  ],
  "meta": { "today": "2026-03-16", "prev_date": "2026-03-09", "days": 7, "count": 1 }
}

Requires daily snapshots to be populated. Returns empty array if rank history not yet available.

GET /api/v1/new-tools

Tools that entered the index for the first time within the last N days.

Parameters

Param	Type	Required	Description
`days`	number	No	Lookback window in days (1–30, default 7)
`limit`	number	No	1–100, default 20
`type`	string	No	`tool` (default), `skill`, or omit for both

Example

curl "https://agentrank-ai.com/api/v1/new-tools?days=7&limit=5"

Response

{
  "tools": [
    {
      "full_name": "acme/new-mcp-tool",
      "tool_type": "tool",
      "rank": 142,
      "score": 41.2,
      "stars": 89,
      "first_seen": "2026-03-14",
      "url": "https://agentrank-ai.com/tool/acme--new-mcp-tool/"
    }
  ],
  "meta": { "today": "2026-03-16", "cutoff": "2026-03-09", "days": 7, "count": 1 }
}

V2 Endpoints — Pro API

Authenticated endpoints with rank history, batch lookups, richer signal data, and no CDN caching. Require a valid API key.

GET /api/v2/search

Same as v1/search with API key auth, rank history hydration, and additional raw fields. Returns your key tier in the response.

Parameters

Param	Type	Required	Description
`q`	string	No	Search query. Omit for full index.
`category`	string	No	`tool`, `skill`, or `agent`. Omit for all.
`sort`	string	No	`score` (default), `rank`, `stars`, `freshness`
`limit`	number	No	1–100, default 20
`offset`	number	No	Pagination offset, default 0
`history`	string	No	Pass `1` to include `rankHistory` on each tool result (last 30 days)

Examples

# curl
curl -H "Authorization: Bearer ark_pro_..." \
  "https://agentrank-ai.com/api/v2/search?q=playwright&history=1"

# JavaScript
const res = await fetch('https://agentrank-ai.com/api/v2/search?q=playwright&history=1', {
  headers: { Authorization: 'Bearer ark_pro_...' }
});

# Python
import httpx
r = httpx.get('https://agentrank-ai.com/api/v2/search',
  params={'q': 'playwright', 'history': '1'},
  headers={'Authorization': 'Bearer ark_pro_...'}
)

Response

{
  "query": "playwright",
  "category": "all",
  "sort": "score",
  "results": [
    {
      "type": "tool",
      "id": "microsoft--playwright-mcp",
      "name": "microsoft/playwright-mcp",
      "description": "Playwright Tools for MCP",
      "score": 78.1,
      "rank": 34,
      "url": "https://agentrank-ai.com/tool/microsoft--playwright-mcp/",
      "raw": {
        "stars": 4200,
        "forks": 312,
        "openIssues": 14,
        "closedIssues": 89,
        "contributors": 28,
        "dependents": 7,
        "lastCommitAt": "2026-03-10T00:00:00.000Z",
        "language": "TypeScript",
        "license": "Apache-2.0",
        "isArchived": false,
        "topics": ["mcp", "playwright"],
        "readmeExcerpt": "...",
        "glamaWeeklyDownloads": 1200,
        "glamaToolCalls": 45000
      },
      "rankHistory": [
        { "date": "2026-03-17", "rank": 34, "score": 78.1 },
        { "date": "2026-03-16", "rank": 36, "score": 77.8 }
      ]
    }
  ],
  "meta": { "total": 1, "limit": 20, "offset": 0 },
  "tier": "pro"
}

No CDN cache — always fresh • Rate limit: per-key daily limit (see tier)

GET /api/v2/tool/{id}

Full signal breakdown for a single tool or skill with rank history (last 30 days). Same ID format as v1/tool.

Examples

# curl
curl -H "Authorization: Bearer ark_pro_..." \
  "https://agentrank-ai.com/api/v2/tool/microsoft--playwright-mcp"

# JavaScript
const res = await fetch('https://agentrank-ai.com/api/v2/tool/microsoft--playwright-mcp', {
  headers: { Authorization: 'Bearer ark_pro_...' }
});

# Python
import httpx
r = httpx.get('https://agentrank-ai.com/api/v2/tool/microsoft--playwright-mcp',
  headers={'Authorization': 'Bearer ark_pro_...'}
)

Response

{
  "type": "tool",
  "id": "microsoft--playwright-mcp",
  "name": "microsoft/playwright-mcp",
  "description": "Playwright Tools for MCP",
  "score": 78.1,
  "rank": 34,
  "githubUrl": "https://github.com/microsoft/playwright-mcp",
  "category": "browser-automation",
  "signals": {
    "stars": 0.862,
    "freshness": 0.943,
    "issueHealth": 0.712,
    "contributors": 0.634,
    "dependents": 0.218,
    "descriptionQuality": 0.7,
    "licenseHealth": 1.0
  },
  "weights": {
    "stars": 0.12, "freshness": 0.23, "issueHealth": 0.24,
    "contributors": 0.09, "dependents": 0.24,
    "descriptionQuality": 0.05, "licenseHealth": 0.03
  },
  "raw": {
    "stars": 4200, "forks": 312, "contributors": 28,
    "openIssues": 14, "closedIssues": 89, "dependents": 7,
    "lastCommitAt": "2026-03-10T00:00:00.000Z",
    "language": "TypeScript", "license": "Apache-2.0",
    "isArchived": false,
    "topics": ["mcp", "playwright", "browser-automation"],
    "matchedQueries": ["browser", "playwright"],
    "readmeExcerpt": "...",
    "glamaWeeklyDownloads": 1200,
    "glamaToolCalls": 45000
  },
  "rankHistory": [
    { "date": "2026-03-17", "rank": 34, "score": 78.1 },
    { "date": "2026-03-16", "rank": 36, "score": 77.8 }
  ]
}

No CDN cache — always fresh • Rate limit: per-key daily limit

POST /api/v2/batch

Batch lookup for up to 50 tools or skills in a single request. Pro and Enterprise only — free tier keys receive a 403.

Request Body

Field	Type	Required	Description
`ids`	string[]	Yes	Array of tool IDs (max 50). Use `owner--repo` for GitHub tools, slug for skills.

Examples

# curl
curl -X POST \
  -H "Authorization: Bearer ark_pro_..." \
  -H "Content-Type: application/json" \
  -d '{"ids": ["microsoft--playwright-mcp", "glama-playwright-mcp-server"]}' \
  "https://agentrank-ai.com/api/v2/batch"

# JavaScript
const res = await fetch('https://agentrank-ai.com/api/v2/batch', {
  method: 'POST',
  headers: {
    Authorization: 'Bearer ark_pro_...',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({ ids: ['microsoft--playwright-mcp', 'glama-playwright-mcp-server'] })
});

# Python
import httpx
r = httpx.post('https://agentrank-ai.com/api/v2/batch',
  json={'ids': ['microsoft--playwright-mcp', 'glama-playwright-mcp-server']},
  headers={'Authorization': 'Bearer ark_pro_...'}
)

Response

{
  "results": {
    "microsoft--playwright-mcp": {
      "type": "tool",
      "id": "microsoft--playwright-mcp",
      "name": "microsoft/playwright-mcp",
      "score": 78.1,
      "rank": 34,
      "githubUrl": "https://github.com/microsoft/playwright-mcp",
      "category": "browser-automation",
      "raw": { "stars": 4200, ... }
    },
    "glama-playwright-mcp-server": {
      "type": "skill",
      "id": "glama-playwright-mcp-server",
      "name": "Playwright MCP Server",
      "score": 82.4,
      "rank": 12,
      "githubUrl": null,
      "category": null,
      "raw": { "stars": null, ... }
    },
    "nonexistent--tool": null
  },
  "count": 2,
  "requested": 3
}

IDs not found return null in results • Max 50 IDs per request • Pro/Enterprise only

Legacy Endpoints

Simpler endpoints without rate limiting. Good for quick lookups and badge embeds.

GET /api/search

Search the index by keyword. Returns both tools and skills.

Parameters

Param	Type	Required	Description
`q`	string	Yes	Search query
`type`	string	No	`tool` or `skill`. Omit to search both.
`limit`	number	No	1–50, default 10

Example

curl "https://agentrank-ai.com/api/search?q=playwright&limit=3"

Response

{
  "query": "playwright",
  "results": [
    {
      "type": "skill",
      "slug": "glama:playwright-mcp-server",
      "name": "Playwright MCP Server",
      "description": "Browser automation and testing via Playwright",
      "score": 82.4,
      "rank": 12,
      "url": "https://agentrank-ai.com/skill/glama-playwright-mcp-server/"
    }
  ]
}

Cache: 5min CDN / 1min browser

GET /api/lookup

Look up a specific entry by its GitHub URL.

Parameters

Param	Type	Required	Description
`url`	string	Yes	GitHub repository URL

Example

curl "https://agentrank-ai.com/api/lookup?url=https://github.com/microsoft/playwright-mcp"

Response (found)

{
  "found": true,
  "type": "skill",
  "slug": "glama:playwright-mcp-server",
  "name": "Playwright MCP Server",
  "score": 82.4,
  "rank": 12
}

Response (not found)

{ "found": false }

GET /api/badge/skill/{slug} /api/badge/tool/{slug}

SVG badge showing the AgentRank score. Embed in READMEs, docs, or websites.

Preview

Markdown

[![AgentRank](https://agentrank-ai.com/api/badge/skill/YOUR-SLUG)](https://agentrank-ai.com/skill/YOUR-SLUG/)

HTML

<a href="https://agentrank-ai.com/skill/YOUR-SLUG/">
  <img src="https://agentrank-ai.com/api/badge/skill/YOUR-SLUG" alt="AgentRank">
</a>

Cache: 24h CDN / 1h browser

Usage Notes

Free endpoints require no authentication. V2 endpoints require an API key.
All JSON endpoints return Access-Control-Allow-Origin: * for browser use.
V1 endpoints enforce a rate limit of 100 requests per minute per IP.
V2 endpoints use per-key daily limits: 100/day (free), 10,000/day (pro), unlimited (enterprise).
V2 responses are never CDN-cached — always reflect live data.
Data updates daily. Scores reflect the latest crawl.
Tool IDs use owner--repo format (replace / with --).
Badge SVGs are CDN-cached — safe to embed anywhere.