# OP Inference

> OP Inference is a paid MCP server for ai models and llm tooling. Call your OP Inference tasks, presets and gallery workflows as tools, billed to your key. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Paid
- **Platforms:** API
- **Website:** https://docs.opinference.com/docs/text/mcp
- **MCP Registry name:** com.opinference/op
- **Categories:** AI models and LLM tooling
- **Listing updated:** 2026-11-16

## About

OP Inference's MCP endpoint turns the things you already have in the platform into tools. Each ready-made task becomes task.<name>, each saved preset becomes preset.<slug>, and a public gallery of agents and workflows can be searched, described and run. A tool call goes through the same route, budget, rate limit and retention rules as a direct API call, and is billed in the same way.

The gallery tools are gallery.search, gallery.describe, gallery.run and gallery.get_run. A run needs a max_cost_usd you choose, and calling run twice starts two billed runs. Listing, search, describe and reading a run are free. Tool annotations mark the free ones read-only, so a client can skip approval, while spending calls ask first.

Clients that support OAuth connect without a key: you sign in, choose a workspace and set a monthly spend cap that defaults to $20, and revoking the app revokes its key within about a minute. The docs warn that tool results and gallery text are untrusted input.

## Features

- **Tasks as tools:** Each ready-made task is callable with its own schema.
- **Preset tools:** Use saved model presets with variables.
- **Gallery search:** Find public workflows and agents by keyword and category.
- **Run with cost cap:** Gallery runs require a max cost you set.
- **OAuth connection:** Sign in and set a monthly spend cap per connected app.
- **Billing parity:** Charges match a direct API call.

## Good for

- Run a structured-extraction task from an assistant
- Reuse a saved prompt preset in an MCP client
- Find and run a public workflow with a cost cap
- Give an agent a key limited to a spend budget

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http op https://api.opinference.com/v1/mcp
```

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

```sh
https://api.opinference.com/v1/mcp
```

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

```json
{
  "mcpServers": {
    "op": {
      "url": "https://api.opinference.com/v1/mcp"
    }
  }
}
```

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

```json
{
  "servers": {
    "op": {
      "type": "http",
      "url": "https://api.opinference.com/v1/mcp"
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "op": {
      "serverUrl": "https://api.opinference.com/v1/mcp"
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "op": {
      "type": "http",
      "url": "https://api.opinference.com/v1/mcp"
    }
  }
}
```

## Tools

- `task.<name>`: Runs a ready-made task with its input schema.
- `preset.<slug>`: Runs a saved preset with a prompt and optional variables.
- `gallery.search`: Searches public workflow and agent listings.
- `gallery.describe`: Describes one gallery listing.
- `gallery.run`: Starts a gallery workflow run with a cost cap.
- `gallery.get_run`: Reads the status of a run.

## FAQ

### How is it billed?

Tool calls are billed exactly as the direct API call would be, to your key.

### Which calls are free?

Listing, ping, discovery, gallery.search, gallery.describe and gallery.get_run.

### What is the default spend cap?

$20.00 a month for an OAuth-connected app, which you can change.

### Can a browser page call it?

No. Requests with an Origin header are refused with 403.

### Is OP Inference free?

No. OP Inference is a paid product.

### How do I add OP Inference to Claude Code?

Run this command in your terminal: claude mcp add --transport http op https://api.opinference.com/v1/mcp

### What are the best alternatives to OP Inference?

The closest alternatives to OP Inference 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/op-inference
