About Together AI MCP by usefulapi
Together AI MCP by usefulapi is a hosted service for Claude, Cursor, and other MCP clients to use Together AI for chat, embeddings, and image generation, and manage fine-tunes and endpoints. It connects through OAuth without a local install.
Key features
- 01List models, files, and dedicated endpoints
- 02Run chat, embeddings, and image generation
- 03Create and cancel fine-tuning jobs
- 04Manage batches and dedicated endpoints
Use cases
- 01Generate chat completions with Together AI models
- 02Create embeddings or images
- 03Submit and track fine-tuning jobs
- 04Create or start and stop dedicated endpoints
MCP support
This hosted MCP server provides MCP clients with Together AI model, file, job, and endpoint queries, along with tools for chat, embeddings, image generation, and resource management. Users provide an API key through OAuth.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://together-ai.usefulapi.io/mcpclaude mcp add --transport http io-usefulapi-together-ai https://together-ai.usefulapi.io/mcp{"mcpServers":{"io-usefulapi-together-ai":{"type":"http","url":"https://together-ai.usefulapi.io/mcp"}}}Open in Cursor{"servers":{"io-usefulapi-together-ai":{"type":"http","url":"https://together-ai.usefulapi.io/mcp"}}}Open in VS Code{"mcpServers":{"io-usefulapi-together-ai":{"type":"http","url":"https://together-ai.usefulapi.io/mcp"}}}{"mcpServers":{"io-usefulapi-together-ai":{"serverUrl":"https://together-ai.usefulapi.io/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
23This tool takes no parameters.
Details
From the project READMETogether AI MCP by usefulapi
Use Together AI from Claude, Cursor, or any MCP client — list models, run chat, embeddings and image generation, and manage fine-tuning jobs, batches and dedicated endpoints. Hosted, no local install: connect with your own Together AI credentials.
Live endpoint: https://together-ai.usefulapi.io/mcp · Homepage: https://usefulapi.io/together-ai
Connect
- Claude (claude.ai, Desktop): open Customize → Connectors, click + Add → Add custom connector, and paste
https://together-ai.usefulapi.io/mcp. - Claude Code:
claude mcp add --transport http together-ai https://together-ai.usefulapi.io/mcp, then run/mcpto log in. - VS Code: Add to VS Code.
- Cursor and other clients: add the URL as a remote MCP server:
{
"mcpServers": {
"together-ai": {
"url": "https://together-ai.usefulapi.io/mcp"
}
}
}
Add only the URL. Do not add an Authorization header or an API key to the client config: the server signs you in with OAuth, and the login page asks for your Together AI credentials.
Step-by-step setup, where to find your credentials, and FAQ: https://usefulapi.io/together-ai/
On first connect you'll provide your Together AI API key (Settings → API Keys). They're validated, stored per-user, and scoped to you — no keys in config files.
Tools
| Tool | Type | What it does |
|---|---|---|
together_whoami | read | Who am I |
together_list_models | read | List models |
together_list_files | read | List files |
together_get_file | read | Get one file |
together_list_fine_tunes | read | List fine-tuning jobs |
together_get_fine_tune | read | Get one fine-tuning job |
together_list_fine_tune_events | read | List a fine-tuning job's events |
together_list_batches | read | List batch jobs |
together_get_batch | read | Get one batch job |
together_list_endpoints | read | List endpoints |
together_get_endpoint | read | Get one endpoint |
together_list_hardware | read | List hardware |
together_list_evaluations | read | List evaluation jobs |
together_chat_completion | write | Chat completion |
together_create_embeddings | write | Create embeddings |
together_generate_image | write | Generate an image |
together_create_fine_tune | write | Create a fine-tuning job |
together_cancel_fine_tune | write | Cancel a fine-tuning job |
together_create_batch | write | Create a batch job |
together_cancel_batch | write | Cancel a batch job |
together_create_endpoint | write | Create a dedicated endpoint |
together_start_endpoint | write | Start a dedicated endpoint |
together_stop_endpoint | write | Stop a dedicated endpoint |
together_ai_usage_status | meta | Usage status (free-tier meter) |
together_ai_request_feature | meta | Request a missing feature |
together_ai_upgrade | meta | Upgrade to Pro (unlimited*) |
together_ai_cancel_subscription | meta | Cancel the Pro subscription |
read tools are read-only; write tools mutate data (clients should confirm them); meta tools report usage, manage your subscription or send a feature request.
Pricing
| Plan | Price | Limit |
|---|---|---|
| Free | $0 | 50 tool calls / month |
| Pro (per user) | $9/mo or $90/yr (2 months free) | Unlimited* |
*Pro has a fair use limit of 10,000 tool calls a month on this server. For more, write to [email protected].
Pro covers this server only. Subscribe with together_ai_upgrade (it returns a Stripe Checkout link). Cancel any time with together_ai_cancel_subscription: Pro continues to the end of the paid period, with no refund for the current period, and running together_ai_upgrade before then undoes the cancel. Or write to [email protected].
License
MIT © usefulapi. Not affiliated with or endorsed by Together AI.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.