slopweaver
A local-first MCP server that helps Claude Code prioritize your work by searching across your tools like GitHub and Slack.
README
SlopWeaver
Help Claude Code answer "what should I work on next?" by searching across your work tools.
Open-source local-first MCP server. BYOK.
Status: pre-alpha. v1.0.0 ships in the coming weeks. Built solo by @lachiejames. Roadmap: tracking issue #2.
What it does
You sit down to work. You ask Claude Code:
"What should I work on next?"
SlopWeaver searches everything — your open PRs, Slack mentions, Linear tickets, threads waiting on your reply, recent activity in repos you care about. Claude synthesizes a priority order and tells you what's worth doing now, what can wait, and what isn't worth your time at all.
You get oriented in 60 seconds instead of 20 minutes of tab-flipping.
Pull-based. Never acts without you. Cognitive partner, not automation tool.
(Demo video lands with v1.0.0.)
Install (coming with v1.0.0)
SlopWeaver runs as a local subprocess of your MCP client over stdio — no HTTP server, no token paste, no auth dance.
Claude Code:
claude mcp add slopweaver -- npx -y @slopweaver/mcp-local
Cursor — add to ~/.cursor/mcp.json (or .cursor/mcp.json in your project):
{
"mcpServers": {
"slopweaver": {
"command": "npx",
"args": ["-y", "@slopweaver/mcp-local"]
}
}
}
Cline — add to ~/.cline/data/settings/cline_mcp_settings.json (or $CLINE_DIR/data/settings/cline_mcp_settings.json if you've set CLINE_DIR):
{
"mcpServers": {
"slopweaver": {
"command": "npx",
"args": ["-y", "@slopweaver/mcp-local"]
}
}
}
Codex CLI — add to ~/.codex/config.toml:
[mcp_servers.slopweaver]
command = "npx"
args = ["-y", "@slopweaver/mcp-local"]
Then connect your work tools (one-time setup). The fastest path is the guided wizard:
npx -y @slopweaver/mcp-local init # detects MCP clients, walks through GitHub + Slack, verifies tokens
Prefer to do it manually?
npx -y @slopweaver/mcp-local connect github # paste a fine-grained PAT (input is hidden)
npx -y @slopweaver/mcp-local connect slack # paste a Slack user token (xoxp-)
(If you'd rather have slopweaver on your PATH directly, npm install -g @slopweaver/mcp-local first, then run slopweaver init.)
Then ask your client: "What should I work on next?" If anything fails, open an issue — a doctor subcommand ships with v1.0.0.
Note: Connecting SlopWeaver to GitHub (so it can poll your PRs and mentions) uses GitHub's own OAuth or a personal access token — that's separate from the MCP transport between your client and SlopWeaver. The MCP layer itself has no auth in v1; stdio inherits the user's trust context.
Try it without connecting anything (demo mode)
If you want to feel the cold-start moment before bringing your own tokens, SlopWeaver ships a demo profile that seeds a separate SQLite database with synthetic GitHub + Slack evidence:
slopweaver demo seed # populates ~/.slopweaver/demo.db with ~22 synthetic rows
slopweaver --demo # runs the MCP server against demo.db instead of slopweaver.db
# (equivalent to SLOPWEAVER_DEMO=1)
slopweaver demo reset # drop + re-seed if the timestamps have drifted
slopweaver demo exit # remove demo.db (restart the server without --demo,
# or unset SLOPWEAVER_DEMO, to return to real mode)
Once the server is running in demo mode, ask your MCP client to call the start_session tool — it serves the synthetic rows exactly the same way it would serve your real data. Only GitHub and Slack are covered (those are the integrations v1.0 actually ships); Linear / Gmail / Calendar are planned for v1.1+ and are intentionally absent from the demo so the snapshot can't claim a richer product than the binary delivers.
The bare slopweaver demo command (with no subcommand) prints a static markdown snapshot to stdout for screenshotting or sharing.
Why local-first
- Your work data stays on your machine. No SaaS, no cloud round-trip, no signup.
- BYOK. Bring your own Anthropic / OpenAI key (only needed for AI features; deterministic context tools work without).
- One binary. SQLite. No Docker, no Postgres, no setup theatre.
Cloud tier (year 2 — coming)
Some things require a server: real-time webhooks instead of polling, mobile push notifications, cross-device sync, always-on observation. The optional SlopWeaver Cloud (launching year 2) adds these. Same code; hosted deployment.
Integrations
v1.0: GitHub + Slack.
v1.1+ (planned): Linear, Gmail, Google Calendar.
Request an integration once issue templates land.
License
MIT.
Security
Integration tokens (GitHub PAT, Slack user token) are stored in the macOS Keychain under the entry slopweaver / <integration> — the local SQLite database holds only presence metadata (slug, account label, timestamps). Audit a stored token with security find-generic-password -a github -s slopweaver -w. On first write the v1 binary is unsigned, so macOS shows a "Keychain Access wants to use the slopweaver entry" prompt; clicking "Always Allow" trusts the binary's path. macOS is the only OS that's QA'd for v1 — Linux Secret Service and Windows Credential Manager work under the hood but are best-effort untested. For vulnerability disclosure see SECURITY.md.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.