Labsco
goperigon logo

rnkflow

from goperigon

Real-time ranked tech feed from Hacker News, arXiv, Dev.to, and more. Query live stories by category, search by keyword, and surface what the tech community is reading right now.

๐Ÿ”ฅ๐Ÿ”ฅFreeQuick setup
<p align="center"> <img src="https://marketing.perigon.io/_next/image?url=%2Flogos%2FLogo-Perigon-Dark.png&w=256&q=75" width="120" alt="Perigon logo" /> </p> <h1 align="center">Perigon&nbsp;MCP&nbsp;Server</h1>

This is the official MCP server for the Perigon news API.

Documentation

For more information on how to use and connect the MCP, visit the MCP docs.

Maintainers

The Perigon MCP Server is developed and maintained by the Perigon engineering team.

Lead Developer: Islem Maboud, responsible for the architecture, implementation, and ongoing development of this MCP server, including the remote transport layer, authentication handling, deployment configuration, and the MCP playground.

Supported tools

The full list of available tools โ€” including names, descriptions, and parameter schemas โ€” is visible in the MCP playground. The tools available to you depend on the scopes granted to your API key.

Signal Insights tools

API keys with the Signal Insights scope unlock an additional set of tools for querying, exporting, and analyzing your AI signals data with a persistent Python sandbox.

Workspace pattern

Signal Insights tools use an explicit workspace handle (per SEP-2567):

  1. Call create_insights_workspace once at the start of a conversation.
  2. Pass the returned workspace ID to every subsequent analysis tool call.
  3. Files written in execute_code or shell persist across calls within the same workspace. Exported data is accessible at /home/user/workspace/artifacts/ inside the sandbox.
  4. If you resume a chat after a restart, the workspace UUID from the prior conversation is still valid โ€” the sandbox kernel will be fresh but your exported S3 artifacts are preserved.

Signal Insights tool list

ToolTypeDescription
signal_insights_create_workspaceSetupCreate a workspace for the conversation. Must be called first.
signal_insights_search_signalsReadSearch signals by name or monitoring objective.
signal_insights_read_signalReadGet signal metadata (schema, event types, counts).
signal_insights_export_eventsDataExport signal events to S3 with optional filters/aggregations. Returns a preview and file path.
signal_insights_execute_codeSandboxExecute Python in a persistent IPython kernel. pandas, numpy, matplotlib and more pre-installed.
signal_insights_shellSandboxRun bash commands in the sandbox.
signal_insights_list_filesFilesList files in the sandbox workspace.
signal_insights_read_fileFilesRead a file from the workspace.
signal_insights_write_fileFilesWrite a file to the workspace.
signal_insights_grepFilesSearch file contents with a regex pattern.
signal_insights_str_replaceFilesFind and replace a string in a file.

Example config โ€” Signal Insights only

{
  "mcpServers": {
    "perigon": {
      "url": "https://mcp.perigon.io/v1/mcp?tools=signal_insights_create_workspace,signal_insights_search_signals,signal_insights_read_signal,signal_insights_export_events,signal_insights_execute_code,signal_insights_shell,signal_insights_list_files,signal_insights_read_file,signal_insights_write_file,signal_insights_grep,signal_insights_str_replace",
      "type": "http",
      "headers": {
        "Authorization": "Bearer YOUR_PERIGON_API_KEY"
      }
    }
  }
}

Example config โ€” Signal Insights combined with news tools

{
  "mcpServers": {
    "perigon": {
      "url": "https://mcp.perigon.io/v1/mcp",
      "type": "http",
      "headers": {
        "Authorization": "Bearer YOUR_PERIGON_API_KEY"
      }
    }
  }
}

With no ?tools= filter, all tools permitted by your API key's scopes are active.

Issues / Contributing

Issues

This MCP server is still in development as we determine what use cases our users want to solve with this server. But if you have any special requests or features you would like to see, don't hesitate to open a github issue on this Repo. We will gladly accept any feedback

Contributing

This tool is intentionally open source so if you want to see some particular feature you open an issue or open a PR and someone at Perigon will review it.

Local development

We are using bun for package mgmt.

Environment Variables

Add the following environment variables to .dev.vars

VariableDescription
ANTHROPIC_API_KEYAnthropic API key (used for playground)
PERIGON_API_KEYPerigon API key (used for playground)
POKEY_SIGNAL_INSIGHTS_BASE_URLInternal Pokey service URL for Signal Insights MCP tools (e.g. http://localhost:3001). Required only when using Signal Insights tools.

If you wish to contribute to the MCP playground (tools inspector & chat), please make sure to modify your network hosts file (/etc/hosts on mac) to include the following

127.0.0.1 local-mcp.perigon.io

This will allow perigon.io cookies to be available for you while doing local development.

# install deps
bun i
# Runs the mcp server and the mcp playground
bun dev