About raportagent-mcp
raportagent-mcp connects MCP clients to RaportAgent for sourced market and compliance research. It creates reports and exposes their content, citations, provenance, and status through the RaportAgent REST API.
Key features
- 01Generate reports from research queries
- 02Select from report templates
- 03Retrieve report status and Markdown
- 04Inspect sources and audit provenance
- 05Cancel reports and list recent reports
- 06Check account plan and credits
Use cases
- 01Map a regulatory landscape
- 02Research a market or competitor
- 03Review citations and report provenance
MCP support
An MCP server that exposes RaportAgent report generation, retrieval, status, source, audit, cancellation, and account operations as tools for MCP clients.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y @raportagent/mcpRAPORTAGENT_API_KEYSecretYour RaportAgent API key (create one at raportagent.com under My account -> API keys)RAPORTAGENT_BASE_URLOverride the API base URL (defaults to https://raportagent.com)Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
6Saved tool list · updated 3w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| depth | enum | No | deep = ~15 min, thorough; fast = ~7 min.deep | fast |
| query | string | Yes | The market/topic to research, e.g. 'EV charging infrastructure Europe 2026'. |
| template | enum | No | Report shape. Use 'compliance' for a legal/regulatory brief.compliance | pitch | saas | ecommerce | realestate | local |
| wait_seconds | integer | No | If >0, poll until the report completes or this many seconds elapse (max 1200). |
Details
From the project READMERaportAgent MCP server
A Model Context Protocol server that puts RaportAgent inside any MCP client — Claude Desktop, Claude Code, Cursor. Ask your assistant to research a market or map a regulatory landscape, and it generates a sourced RaportAgent report, then pulls the finished markdown and its audit trail.
Thin wrapper over the RaportAgent REST API (/v1), so it inherits per-key rate limits, the
credit model, and the audit trail unchanged.
Tools
| Tool | What it does |
|---|---|
generate_report | Start a report (query, optional template: compliance/pitch/saas/ecommerce/realestate/local/battlecard/duediligence/fintech). Returns a report_id; optionally polls with wait_seconds. Costs 1 credit (2 for battlecard). |
get_report_status | queued / running / completed / failed |
get_report | The finished report's full markdown + sections |
get_report_audit | Provenance: AI models, agents, source counts, SHA-256 of the exact content |
get_report_sources | Every cited source with link-health counts (working / uncertain / dead) |
cancel_report | Cancel a queued/in-progress report and refund its credit |
list_reports | Your recent reports |
get_account | Remaining credits and plan |
Setup
- Get an API key: RaportAgent → My account → API keys (an
ra_live_…key). - Register it with your MCP client — no clone, no build,
npxpulls the published package.
Claude Desktop
Add to claude_desktop_config.json
(macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):
{
"mcpServers": {
"raportagent": {
"command": "npx",
"args": ["-y", "@raportagent/mcp"],
"env": {
"RAPORTAGENT_API_KEY": "ra_live_xxx"
}
}
}
}
Claude Code
claude mcp add raportagent -e RAPORTAGENT_API_KEY=ra_live_xxx -- npx -y @raportagent/mcp
Cursor
Same command/args/env shape as Claude Desktop above, under Cursor's MCP settings.
Running from source (contributing / debugging only)
git clone https://github.com/raportagent/raportagent-mcp.git
cd mcp-server
npm install
npm run build
Then point command/args at node / dist/index.js with an absolute path instead of npx.
Config
| Env | Default | Notes |
|---|---|---|
RAPORTAGENT_API_KEY | — | Required. ra_live_… key. |
RAPORTAGENT_BASE_URL | https://raportagent.com | Override for self-hosted / staging. |
Notes
- Generation typically takes ~15 minutes.
generate_reportreturns immediately with areport_id; the assistant should pollget_report_status(or passwait_seconds) beforeget_report. stdoutis the MCP transport — the server logs only tostderr.- The API key is only ever sent to your configured
RAPORTAGENT_BASE_URL.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.