InnerVault

driftnet

Stale Official MCP Registry, io.github.KevinChu1110

A local catalogue of the links your chats share: search, save and file them. Any model, or none.

Open the source (opens github.com) https://github.com/kevinchu1110/driftnet

Before you use it

  • Runs on your computer with your permissions. Read what it can do before connecting it to an agent.
  • Needs credentials: AWS_ACCESS_KEY_ID, AWS_BEARER_TOKEN_BEDROCK, AWS_SECRET_ACCESS_KEY, DRIFTNET_LLM_API_KEY. Use a key with the least access you can.

What it is

MCP servers from io.github.KevinChu1110.

The problem it solves and who it suits haven't been reviewed by InnerVault. The description above is the source's own.

How to get it

npm package (stdio)

driftnet@1.2.1

From the server's entry in the official MCP Registry

Settings it asks for

  • DRIFTNET_HOME · Where settings and the catalogue live (default ~/.driftnet)
  • DRIFTNET_LLM_PROVIDER · anthropic, openai, gemini, bedrock, vertex, claude-cli or gemini-cli; empty for no summaries
  • DRIFTNET_LLM_MODEL · Model name for that provider
  • DRIFTNET_LLM_API_KEY secret · API key for that provider (gemini also reads GEMINI_API_KEY; for vertex it may be a raw access token)
  • DRIFTNET_LLM_BASE_URL · OpenAI-compatible endpoint for the openai provider (Ollama, OpenRouter, Gemini…)
  • DRIFTNET_LLM_REGION · Region for the bedrock provider (or AWS_REGION) and the vertex provider (default us-central1)
  • AWS_ACCESS_KEY_ID secret · bedrock provider: AWS access key ID (with AWS_SECRET_ACCESS_KEY, and AWS_SESSION_TOKEN for temporary credentials)
  • AWS_SECRET_ACCESS_KEY secret · bedrock provider: AWS secret access key
  • AWS_BEARER_TOKEN_BEDROCK secret · bedrock provider: a Bedrock API key, used instead of the AWS keys
  • GOOGLE_APPLICATION_CREDENTIALS · vertex provider: path to a service-account JSON key file (or DRIFTNET_VERTEX_CREDENTIALS)
  • GOOGLE_CLOUD_PROJECT · vertex provider: Google Cloud project id (optional when the key file names one)
  • DRIFTNET_JUDGE · Who files each link before the model: jev, rules (free, no model), llm, heuristic, off, or a plugin's judge name

Related

  1. GitHub MCP server

    MCP servers

    GitHub's official MCP server. It lets an agent read and work with repositories, issues and pull requests.

    Hand-checked InnerVault (hand-checked), GitHub 34.3k ★MIT

    https://api.githubcopilot.com/mcp/
  2. Apify MCP server

    MCP servers

    Lets an agent run Apify's ready-made scrapers to pull data from websites, search engines, maps and social platforms.

    Hand-checked InnerVault (hand-checked), Apify 16.0k ★MIT

  3. Official MCP Registry

    MCP servers

    The MCP project's own registry of public MCP servers, so you can look for a connector before building one.

    Hand-checked InnerVault (hand-checked), MCP project Moving from MIT to Apache-2.0

  4. MCP reference servers

    MCP servers

    Small example servers from the MCP project (filesystem, git, fetch, memory and more) that show how MCP works.

    Hand-checked InnerVault (hand-checked), MCP project Moving from MIT to Apache-2.0

  5. Anthropic's guide to adding MCP servers to Claude Code, with examples of local and remote servers.

    Hand-checked InnerVault (hand-checked), Anthropic Anthropic documentation

  6. Playwright MCP

    MCP servers

    Lets an agent drive a real browser: open pages, click, fill in forms and read what's on screen.

    Hand-checked InnerVault (hand-checked), Microsoft Apache-2.0

    claude mcp add playwright npx @playwright/mcp@latest