# MCP

All Synth MCP tools support both 1-hour and 24-hour forecast horizons. When using tools like `get-prediction-percentiles`, `get-insights-volatility`, or `get-insights-option-pricing`, you can specify the horizon by passing `horizon=1h` or `horizon=24h` (default).

MCP allows AI agents to integrate seamlessly to external data sources and tools.

Synth API MCP is a remote MCP server at <https://mcp.synthdata.co/v1/mcp/apikey/json>.

There are 2 authentications methods:

* OAuth (coming soon)
* API Key

To authenticate the MCP with your API key, you need to add custom headers:

* header key: Authorization
* header value: Bearer \<your synth api key>

### Connect with Claude Desktop

Download the Desktop Extension: [synth.mcpb](https://github.com/mode-network/synth-mcp/releases/download/v0.0.10/synth.mcpb).&#x20;

To add the extension double click on the file and Claude Desktop will ask you to install, and input your API key.

### Connect with Claude Code CLI

Claude Code provides a CLI to use the agentic coding tool of Anthropic. Read more [here](https://code.claude.com/docs/en/overview#terminal).

To add the SynthData MCP server, type the command:

```shellscript
claude mcp add --transport http synthdata https://mcp.synthdata.co/v1/mcp/apikey/json --header "Authorization: Bearer <your api key>"
```

### Connect with mcp-remote

mcp-remote is a tool that allows you to install a remote MCP server for a local MCP client. It creates a local STDIO MCP to proxies to the remote HTTP Stream Synth MCP. Read more [here](https://www.npmjs.com/package/mcp-remote).

Type the command to add the Synth MCP:

```shellscript
npx -y mcp-remote https://mcp.synthdata.co/v1/mcp/apikey/json --header "Authorization: Bearer <your api key>" --allow-http
```

### Connect with Vercel AI SDK

Vercel AI SDK supports connecting to remote MCP servers. Read more [here](https://ai-sdk.dev/docs/ai-sdk-core/mcp-tools).

```typescript
  const mcpClient = await createMCPClient({
    transport: {
      type: "http",
      url: "https://mcp.synthdata.co/v1/mcp/apikey/json",

      headers: { Authorization: `Bearer ${process.env.SYNTH_API_KEY}` },
    },
  });

  const tools = await mcpClient.tools();

  const result = streamText({
    model: gateway(modelId),
    tools,
    system: "You are an agent working with SynthData's MCP.",
    messages: convertToModelMessages(messages),
    onError: (e) => {
      console.error("Error while streaming.", e);
    },
  });
```

### Connect with OpenAI

OpenAI platform allows you to connect to MCP server:

1. go to the Open AI Platform: <https://platform.openai.com/chat>
2. on the tools configuration, click "Add"
3. chose "MCP server"
4. click on "+ Server" button
5. input the url: <https://mcp.synthdata.co/v1/mcp/apikey/json>
6. chose a label: synth
7. on authentication, select "Custom headers"
8. type "Authorization" in the header input
9. type "Bearer \<your api key>" in the value input, change with your API key that you create here [https://dashboard.synthdata.co/api-keys](https://dashboard.synthdata.co/api-keys/)
10. Click on "Connect"

You can now prompt the LLM and ask for Synth Insights.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.synthdata.co/mcp.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.