Rhumb
MCP server that tells a coding agent which AI model to call and what to connect
Answered our latest tool-list check 3 hours ago with 9 tools. It will be marked verified after 3 successful checks in 7 days.
- Pricing
- Freemium
- Platforms
- API, CLI
- Runs
- Remotely
- Version
- 1.0.0
- Listing updated
- 2 days ago
About Rhumb
Rhumb is a freemium MCP server for ai models and llm tooling. It runs as a hosted remote server, so there is nothing to install locally.
A coding agent that writes a model id from memory tends to write an outdated one. Rhumb answers from a catalogue instead: you describe a need, such as cheap summaries on the providers you already use, and it returns one model, its id on that provider, its price and the reason.
The docs describe two parts that work independently. The remote MCP server at rhumbdata.com/v1/mcp exposes the API as tools and takes the same API key in an Authorization header. An optional rhumb skill file tells Claude Code when to ask and what never to send. The agent is meant to find your providers from environment variable names and SDKs, never key values.
A separate stack tool covers the systems an agent works with, answering each one's own MCP server or command-line tool, a model, warnings and a starting config. Rhumb says it is not a deployment recommender and does not compare agent runtimes.
Install
Run in a terminal
claude mcp add --transport http rhumb https://rhumbdata.com/v1/mcpAdd to Settings → Connectors → Add custom connector
https://rhumbdata.com/v1/mcpAdd to ~/.cursor/mcp.json
{
"mcpServers": {
"rhumb": {
"url": "https://rhumbdata.com/v1/mcp"
}
}
}Add to .vscode/mcp.json
{
"servers": {
"rhumb": {
"type": "http",
"url": "https://rhumbdata.com/v1/mcp"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"rhumb": {
"serverUrl": "https://rhumbdata.com/v1/mcp"
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"rhumb": {
"type": "http",
"url": "https://rhumbdata.com/v1/mcp"
}
}
}Tools this server provides (9)
Read directly from the server.
- resolve
- Which model to call for a need, on the providers the user has accounts with: the provider's own model id, its price, capabilities and the reason. Give `providers` and a `task` or a `purpose`; add `needs`, `budget`, `origin` and other constraints to narrow it. `shape` asks for a team or an escalation ladder instead of one model. Counts as one call. An answer of `no_candidate` names what ruled every
- plan
- Reads a description of the work (a prompt, tags, attachment types and sizes) into a need, then resolves it as `resolve` does. Use `resolve` when you can state the task or purpose yourself; use this when you only have the user's words. Counts as more than one call. The answer depends on these goals; have the ones that matter in hand from the plan, the repository or the user before calling: cost (`o
- stack
- For an agent that has to work with named systems (Linear, GitHub, Postgres…): what to connect for each, and which model to call. Give `systems`, `target` (local or cloud) and the `harness` the agent runs in; add `model.providers` to get a model too. Each system's choice is its owner's own MCP server or command-line tool, or none, picked by fixed rules from public records and not reviewed; `warning
- platform
- Where an app's agent runs: give `requires` (baa, gpu, private_network, edge, long_running, websockets, scheduled_jobs, large_uploads) with `longest_turn_seconds` and `largest_upload_mb` where they apply, and the goals you have. The answer is `major` (a platform is warranted: a requirement no curated runtime meets, a model provider it hosts, or an account held; name its `system` in a `stack` call n
- list_models
- The catalog's models with each provider's offering, price, capabilities and quality scores, filtered by task, providers and needs. Returns 20 models unless `limit` says otherwise; page with `offset`. Counts as one call.
- get_model
- One model by id (`vendor/model`, e.g. `anthropic/claude-sonnet-4.5`): every provider's offering with its price history, capabilities, limits and dates, and the model's quality scores. Counts as one call.
- list_purposes
- The purpose presets `resolve` accepts: named jobs such as reading receipts or a coding agent, each with the task, needs and budget it stands for. Free.
- list_goals
- The goals an answer depends on (cost, depth, where a model's maker is based, where data may stay, languages, the kind of data, the accounts already held), each with one question to ask the user, the options to offer and the parameters each answer is sent as, and the rules for asking. Call it once when planning an app. Free.
- get_vocabulary
- Every name the other tools accept: tasks, capabilities and their levels, providers, origin groups, budgets and team strategies. Free.
What it does
- Model resolution
- resolve returns one model, a team or an escalation ladder for a need on your providers.
- Plan from a prompt
- plan reads a description of the work into a need, then resolves it.
- Stack recommendations
- stack suggests each system's MCP server or CLI, a model, warnings and a .mcp.json.
- Model catalogue
- list_models and get_model show offerings, prices, price history and scores.
- Claude Code skill
- A one-file skill tells the agent when to ask and what not to send.
- Per-key controls
- Keys carry scopes, IP allowlists and a monthly cap; test keys do not use quota.
Good for
- Pick a cheap summarisation model for the providers a project already uses
- Look up a model's id and price on OpenRouter before writing it into code
- Choose servers and tools for an agent that connects to Linear and GitHub
- Compare price history for a model across providers
- Get a starting .mcp.json for an agent's systems
Pricing
Freemium, from $9 a month
Free 1,000 calls/month; Builder $9, Pro $39, Scale $149 a month
Prices change. Check rhumbdata.com before you buy.
Rhumb FAQ
How do I add it to Claude Code?
Create a free key, then run claude mcp add --transport http rhumb https://rhumbdata.com/v1/mcp with an Authorization Bearer header.
What does it send to Rhumb?
Provider names, a task or purpose and constraints; for plan also tags and a prompt. Never provider keys or files.
Is a prompt stored?
The docs say a prompt sent to plan is never stored or logged, and can be read on Rhumb's server only if you set allow_remote to false.
How are calls counted?
Each tool call is the API call it stands for; plan counts as 5, and list_purposes and get_vocabulary are not counted.
Does it say where to deploy my app?
No. It does not recommend clouds or hosting platforms.
Is Rhumb free?
Rhumb has a free plan, with paid plans for more usage or features. Paid plans start at $9 a month.
How do I add Rhumb to Claude Code?
Run this command in your terminal: claude mcp add --transport http rhumb https://rhumbdata.com/v1/mcp
What are the best alternatives to Rhumb?
The closest alternatives to Rhumb are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.
Alternatives to Rhumb

REA
MCP serverCLI and MCP server that lets coding agents reverse engineer apps and binaries
Runs locally- Open source
- 464K GitHub stars
- Developer tools

Codebase Memory
MCP serverLocal MCP server that builds a code knowledge graph for AI coding agents
Runs locally- Open source
- 48.7K GitHub stars
- Knowledge and memory

GitHub
MCP serverOfficial MCP server connecting AI assistants to GitHub repos, issues and workflows
Not checked yet- Remote
- Open source
- 34.5K GitHub stars
- Developer tools

Serena MCP: the IDE for your agent
MCP serverMCP toolkit giving coding agents symbol-level code retrieval and editing
Runs locally- Open source
- 31.5K GitHub stars
- Developer tools

XERJ
MCP serverLocal search engine for AI agents with cited answers over your code and docs
Runs locally- Open source
- 19.2K GitHub stars
- Knowledge and memory

Ouroboros
MCP serverSpec-first layer that pins requirements and verifies results of AI coding agent work
Runs locally- Open source
- 6.4K GitHub stars
- Developer tools