Provides search and execute tools over a unified catalog of Stellar ecosystem services and skills, enabling agents to discover and invoke Stellar-related capabilities.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Stellar Raven find a Stellar ecosystem service for swaps and execute it"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Stellar Raven is a remote MCP server on Cloudflare Workers. It gives agents two tools over one catalog of Stellar ecosystem sources: Lumenloop, Stellar Light/Scout, Stellar Docs, and selected ecosystem skills.
search finds service operations and skills.
execute runs agent-written JavaScript in a Dynamic Worker isolate with no network access.
Host-side adapters make every service call and hold every secret.
The server instructions also include a generated source-family map. Agents use it to plan which source family can answer a question.
Source family | Run by | How Raven reads it |
Lumenloop ( | The independent Lumen Loop team (lumenloop.com) | Its API, with an API key that Lumen Loop issued to Raven |
Stellar Light/Scout ( | The Stellar Light team (stellarlight.xyz) | Its read-only public API, which it publishes for AI tools and agents |
Stellar Docs ( | The Stellar Development Foundation (developers.stellar.org) | The search index of the official docs site |
Selected ecosystem skills come from Lumen Loop, OpenZeppelin, the Stellar Development Foundation,
Stellar Light, and Trustless Work. Raven pins each source at a reviewed commit, and each skill keeps
its own license. ecosystem-skills/README.md describes the pin set.
Related MCP server: SIEMen
Add https://raven.stellar.org/mcp to an MCP client that supports streamable HTTP and OAuth.
Raven is its own OAuth authorization server, and WorkOS AuthKit handles sign-in. Compatible
clients find and complete this flow automatically. An access token lasts 1 hour. Clients refresh
it automatically for 90 days, then the user signs in again.
Page | URL |
User guide and troubleshooting | |
Browser Playground | |
Terms and privacy | |
Health |
For help, use the #raven channel in the
Stellar Developers Discord. To report a vulnerability, read
SECURITY.md.
Use Node 24 (.nvmrc) and run npm ci.
Create .dev.vars with the variable names in the .dev.vars step of
ci.yml. Set DEV_ALLOW_UNAUTHENTICATED=true to skip OAuth.
This bypass works only for the loopback hosts localhost, 127.0.0.1, ::1, and [::1].
A Lumenloop call needs a real LUMENLOOP_API_KEY. A Stellar Docs call needs real
ALGOLIA_APPLICATION_ID_DOCS and ALGOLIA_API_KEY_DOCS values. Without them, the call returns
an error. A Stellar Light/Scout call needs no key.
Run npm run typegen. It generates env.d.ts from wrangler.jsonc and the names in
.dev.vars.
Run npm run dev and connect a client to http://localhost:8787/mcp. Restart the server after
you edit .dev.vars.
The Worker reads .dev.vars. Maintenance scripts, such as scripts/refresh-inventory.mjs, read
.env. Git ignores both files.
npm run typecheck # tsc
npm test # unit and contract tests (offline)
npm run test:smoke # the assembled Worker and the Dynamic Worker boundary
npm run build # dry-run the Worker bundletest/README.md maps the test suites.
Question | Document |
What is in scope, and which decisions are open? | |
How do | |
How do I deploy and operate Raven? | |
How is usage counted and retained? | |
How do I contribute? | |
How do agents work in this repository? | |
How do I report a vulnerability? |
Apache-2.0 © 2026 Stellar Development Foundation. Third-party code, fonts, and upstream skill content keep their own licenses. Read THIRD-PARTY-NOTICES.md.
This server cannot be deployed
AI agent skills marketplace — token-efficient skill search & execution
The governed runtime for agent skills. Search the catalog and inspect a skill before running it.
One place to build, share, and govern the skills and tools your AI agents use at work.
Discover, compare, route, and execute machine-accessible capabilities for AI agents.
