Saptiva

You can run the MCP-Saptiva server locally to interact with Saptiva AI models and tools from your own environment. This server exposes chat, reasoning, OCR, embeddings, and a built-in help system, letting you compose complex AI workflows with ease and flexibility.

How to use

Start the MCP server locally and connect to it from your MCP client or application. You can access chat completions with Saptiva Turbo or Cortex models, perform reasoning tasks with Cortex, extract text from images using OCR, generate semantic embeddings, and invoke the built‑in help system for guidance and examples. The server surfaces endpoints and tools you can call from your client, enabling you to build end‑to‑end AI assistant experiences, data extraction pipelines, and model-driven workflows.

How to install

Prerequisites: ensure you have Node.js installed on your system. You will also need a Saptiva API key to authenticate requests.

1. Install dependencies and build the server.

npm install
npm run build

Configuration and startup

Configure the server to use your Saptiva API key and run it locally. The server can be started via a local stdio configuration that runs the built JavaScript entry point and passes the API key through environment variables.

2. Prepare environment variables and start the server using a local runtime.

Usage with Claude Desktop configuration

If you want to connect from Claude Desktop, add a MCP server entry that runs the local MCP server and passes your API key.

{
  "mcpServers": {
    "saptiva": {
      "command": "node",
      "args": ["/path/to/MCP-Saptiva/dist/index.js"],
      "env": {
        "SAPTIVA_API_KEY": "your_api_key_here"
      }
    }
  }
}

What you can do with the available tools

You have access to a set of tools that let you chat with Saptiva models, perform advanced reasoning, extract text from images, generate embeddings, and list models or pricing. Use the built‑in help system to discover topics and examples for quick starts, chat scenarios, reasoning workflows, OCR tasks, and embedding generation.

Examples of typical tasks

• Chat: ask questions to Saptiva Turbo or Cortex models and continue the conversation across turns.

• Reasoning: perform chain‑of‑thought reasoning for complex questions or multi‑step problems.

• OCR: extract text from documents or images to enable searchable content and downstream processing.

• Embeddings: convert text into semantic vectors for similarity search and retrieval.

Security and keys

Protect your API key. Store it in environment variables and avoid exposing it in client applications or logs.

Troubleshooting notes

If you encounter authentication issues, verify that SAPTIVA_API_KEY is correctly set in the environment of the running process and that the key is active in your Saptiva account.

Notes

The server exposes tooling for model access, reasoning, OCR, embeddings, model listings, and built‑in help to guide usage and examples.

Development

Install dependencies, build, and run tests as part of your development workflow.

# Install dependencies
npm install

# Build the project
npm run build

# Run tests
node test-comprehensive.mjs

Available tools

saptiva_chat

Send chat completion requests to Saptiva AI models and receive model responses.

saptiva_reason

Perform complex reasoning with chain-of-thought using the Cortex model.

saptiva_ocr

Extract text from images using the vision model.

saptiva_embed

Generate semantic embeddings for text to enable semantic search and similarity.

saptiva_batch_embed

Generate embeddings for multiple texts in a batch.

saptiva_list_models

List all available models with pricing and capabilities.

saptiva_help

Access built-in documentation and examples to help you use the server.