
StackMap
io.github.Ship-With-AIPublished by GitHub account Ship-With-AIv0.2.2Updated Oct 24, 2026
Search 344 human-reviewed AI/agent repos and the 817 typed connections between them.
Overview
Lets an assistant search a curated, human-reviewed catalog of open-source AI and agent repositories and the typed connections between them.
- What you can do
- StackMap exposes a curated map of the open-source AI stack: 344 repositories across 21 topics and 817 typed connections, each labeled complements, alternative, or built_with and reviewed by a human. Tools include search_catalog for topic queries, get_repo for a single repository including when not to use it, get_alternatives for replacements, continue_stack for what pairs well with a given set of tools, and build_stack for a starter stack from a described goal. Answers carry the curator's written reasoning rather than a list of links.
- When to use it
- Useful when choosing between AI or agent frameworks, libraries, and tools, or when assembling a stack for a described goal such as a local-first RAG pipeline. Also helpful for finding alternatives to a tool already in use and for seeing which components are commonly combined.
- Requirements
- Runs locally as a stdio process, installed via npx from the npm package @ship-with-ai/stackmap-mcp, so Node.js and network access are needed. No authentication, no environment variables, and no headers are declared. It reads the site's static JSON API, and the catalog is date-versioned.
Installation
In SourceWeft
- Open StackMap in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
@ship-with-ai/stackmap-mcp
Query StackMap — the curated map of the open-source AI stack — from any MCP client. 344 repos across 21 topics, 817 connections, every one typed (complements / alternative / built_with) and human-reviewed, with the why written down — including when NOT to use a tool.
No auth. No state. No telemetry. Reads the site's static JSON API; the catalog is date-versioned (v2026.07.24 style) so you always know how fresh the map is.
Install
Claude Code
Cursor — .cursor/mcp.json:
Codex CLI — ~/.codex/config.toml:
Tools
Every answer carries the curator's written why — the sentence a colleague would tell you over your shoulder, not a link dump.
How the data is made
Nothing on the map auto-publishes: an LLM proposes, a human approves, sharpens, or refuses — and both the refusal ledger and the dismissal ledger are public. Suggest a repo at /suggest; dispute any edge from its repo page.
MIT · by ShipWithAI
Source: README.md at commit e9179e8
Tools
0Version history
1- v0.2.2LatestSep 16, 2026
Score
50% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts