# OpenNodes

> OpenNodes is an open-source MCP server for ai models and llm tooling. Pick and price AI inference nodes and models from a registry that measures them. It can run locally on your machine or as a hosted remote server.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI, API, self-hosted
- **Website:** https://opennodes.io
- **Repository:** https://github.com/opennodes-io/opennodes/tree/HEAD/impl/packages/cli
- **MCP Registry name:** io.github.opennodes-io/opennodes
- **Categories:** AI models and LLM tooling
- **Listing updated:** 2026-11-15

## About

OpenNodes keeps a registry of AI inference endpoints and checks what each one really does: which model it serves, how much context works in a live test, time to first token, tokens per second, and where it runs. Its homepage snapshot lists 368 sites in 38 countries, and says imported listings show only claims until their operators register and are verified.

The MCP server exposes seven read-only tools. recommend chooses a node and model for a task with a price and an explanation; the prompt is analysed on your machine and only extracted features are sent. For models you run yourself it also estimates the memory each size needs and leaves out what will not fit. Other tools search offerings, vet a provider, estimate chain costs, and look up where models run and what residency commitments providers make.

It can run as a stdio server from the npm package @opennodes/cli or as a remote endpoint. The same CLI includes a local OpenAI-compatible gateway with virtual models such as onp/auto and onp/auto-eu, failover across ranked nodes, and signed usage receipts.

## Features

- **Task-based recommendation:** Suggests the best node and model for a task, priced and explained.
- **Measured node data:** Reports checked model identity, working context window, speed and location.
- **Memory-aware local picks:** Estimates memory per model size and skips sizes that cannot fit.
- **Residency lookup:** Finds where models run and what providers promise about data location.
- **Local gateway:** OpenAI-compatible server that routes requests across nodes with failover.
- **Remote or stdio:** Use the hosted endpoint or run the npm package locally.

## Good for

- Choose a model and provider for a specific job with a price estimate
- Vet an inference provider before sending it traffic
- Find EU-hosted endpoints for a data-residency requirement
- Route OpenAI-SDK apps across local and hosted models

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http opennodes https://registry.opennodes.io/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://registry.opennodes.io/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "opennodes": {
      "url": "https://registry.opennodes.io/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "opennodes": {
      "type": "http",
      "url": "https://registry.opennodes.io/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "opennodes": {
      "serverUrl": "https://registry.opennodes.io/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "opennodes": {
      "type": "http",
      "url": "https://registry.opennodes.io/mcp"
    }
  }
}
```

## Tools

- `recommend`: Best node and model for a task, priced and explained.
- `search_offerings`: Search inference offerings.
- `get_offering`: Details of one offering.
- `get_node`: Vet a provider: operator, tier history, availability.
- `estimate`: Pre-price a chain of steps.
- `find_locations`: Where models run.
- `find_residency`: What providers promise about where data stays.

## FAQ

### Does OpenNodes see my prompts?

The site says it is never in the data path, and recommend sends only extracted features, not the prompt.

### What is the install command?

The stdio server comes from npx @opennodes/cli mcp; a remote endpoint is at registry.opennodes.io/mcp.

### Does it need an account?

The Ollama router is described as needing no account or API key.

### What Node version is needed?

The router needs Node.js 22.5 or later.

### Is OpenNodes free?

Yes. OpenNodes is open source, so you can use and self-host it for free.

### How do I add OpenNodes to Claude Code?

Run this command in your terminal: claude mcp add --transport http opennodes https://registry.opennodes.io/mcp

### What are the best alternatives to OpenNodes?

The closest alternatives to OpenNodes are Claude Code Ultimate Guide, signafAIde, Toolkien, Lugh and HypePaper.

## Alternatives

- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [signafAIde](https://ai-powered.com/mcp/signafaide): signafAIde is a freemium MCP server for security.
- [Toolkien](https://ai-powered.com/mcp/toolkien): Toolkien is an open-source MCP server for ai models and llm tooling.
- [Lugh](https://ai-powered.com/mcp/lugh): Lugh is an open-source MCP server for ai models and llm tooling.
- [HypePaper](https://ai-powered.com/mcp/hypepaper): HypePaper is a freemium MCP server for ai models and llm tooling.
- [Agent Data Tools: Compact (find, describe, run)](https://ai-powered.com/mcp/agent-data-tools-compact-find-describe-run): Agent Data Tools: Compact (find, describe, run) is a freemium MCP server for developer tools.
- [Angel Agent Guard: pre-purchase check + prompt-injection scan](https://ai-powered.com/mcp/angel-agent-guard-pre-purchase-check-prompt-injection-scan): Angel Agent Guard: pre-purchase check + prompt-injection scan is a freemium MCP server for ai models and llm tooling.
- [Mizan Calibration](https://ai-powered.com/mcp/mizan-calibration): Mizan Calibration is a free MCP server for ai models and llm tooling.

---
Source: https://ai-powered.com/mcp/opennodes
