Bring Llm
to Claude Desktop
Learn how to connect Cohere to Claude Desktop and start using 6 AI agent tools in minutes. Fully managed, enterprise secure, and ready to use without writing a single line of code.
GDPR Free for Subscribers
What is the Cohere MCP Server?
Connect your Cohere account to any AI agent and leverage enterprise-grade AI models through natural conversation.
What you can do
- Model Discovery — List all available Cohere models with their names, capabilities and context lengths
- Chat API — Send conversations to Command models (command-r-plus, command-r, command-r7b) and receive responses with citations and tool call support
- Embeddings — Generate vector embeddings for semantic search with multiple embedding types (float, int8, uint8, binary)
- Reranking — Rerank documents by relevance to a search query using Cohere's industry-leading reranking models
- Tokenization — Tokenize and detokenize text for estimating token counts and debugging
How it works
1. Subscribe to this server
2. Enter your Cohere API Key
3. Start using Cohere models from Claude, Cursor, or any MCP-compatible client
No more switching between API tools to interact with Cohere. Your AI acts as an LLM orchestration layer.
Who is this for?
- Developers — quickly send messages to Command models, generate embeddings and rerank search results without writing HTTP code
- ML Engineers — discover available models, compare capabilities and generate embeddings with multiple types (float, int8, binary)
- Search Teams — rerank documents by relevance, tokenize text and generate embeddings for search index building
Built-in capabilities (6)
Requires the model ID (e.g. "command-r-plus", "command-r", "command-r7b") and messages array in JSON format. Each message must have a "role" ("user", "assistant", "system" or "tool") and "content" (text or array of content blocks). Optionally set max_tokens, temperature (0-1), p (nucleus sampling 0-1) and tools array for function calling. Returns the model's response with text, citations and tool calls. Send a chat message to a Cohere model
Requires the token IDs array. Returns the reconstructed text. Useful for debugging and verifying tokenization. Detokenize token IDs back to text using Cohere
Requires the model ID (e.g. "embed-v4", "embed-v3"), texts array and input_type ("search_document", "search_query", "classification", "clustering"). Returns embedding vectors for each input text. Useful for semantic search, similarity comparison and vector database storage. Generate embeddings using Cohere
Each model returns its name (e.g. "command-r-plus", "command-r", "embed-v4", "rerank-v3.5"), endpoint compatibility, context length and tokenization info. Use this to discover which models are available and their capabilities. List all available Cohere models
Requires the model ID (e.g. "rerank-v3.5", "rerank-english-v3.0"), query text and documents array. Optionally set top_n to return only the top N results. Returns ranked documents with relevance scores. Rerank documents by relevance to a query
Requires the text to tokenize and optionally the model. Returns the list of token IDs and token strings. Useful for estimating token counts before sending to chat or embed endpoints. Tokenize text using Cohere
Why Claude Desktop?
Claude Desktop is the definitive way to connect Cohere to your AI workflow. Add Vinkius Edge URL to your config, restart the app, and Claude immediately exposes all 6 tools in the chat interface. ask a question, Claude calls the right tool, and you see the answer. Zero code, zero context switching.
- —
Claude Desktop is the reference MCP client. it was designed alongside the protocol itself, ensuring the most complete and stable MCP implementation available
- —
Zero-code configuration: add a server URL to a JSON file and Claude instantly discovers and exposes all available tools in the chat interface
- —
Claude's extended thinking capability lets it reason through multi-step tool usage, chaining multiple API calls to answer complex questions
- —
Enterprise-grade security with local config storage. your tokens never leave your machine, and connections go directly to Vinkius Edge network
Cohere in Claude Desktop
Cohere and 3,400+ other MCP servers. One platform. One governance layer.
Teams that connect Cohere to Claude Desktop through Vinkius don't need to source, host, or maintain individual MCP servers. Every tool call runs inside a hardened runtime with credential isolation, DLP, and a signed audit chain.
Raw MCP | Vinkius | |
|---|---|---|
| Server catalog | Find and host yourself | 3,400+ managed |
| Infrastructure | Self-hosted | Sandboxed V8 isolates |
| Credential handling | Plaintext in config | Vault + runtime injection |
| Data loss prevention | None | Configurable DLP policies |
| Kill switch | None | Global instant shutdown |
| Financial circuit breakers | None | Per-server limits + alerts |
| Audit trail | None | Ed25519 signed logs |
| SIEM log streaming | None | Splunk, Datadog, Webhook |
| Honeytokens | None | Canary alerts on leak |
| Custom domains | Not applicable | DNS challenge verified |
| GDPR compliance | Manual effort | Automated purge + export |
Why teams choose Vinkius for Cohere in Claude Desktop
The Cohere MCP Server runs on Vinkius-managed infrastructure inside AWS — a purpose-built runtime with per-request V8 isolates, Ed25519 signed audit chains, and sub-40ms cold starts. All 6 tools execute in hardened sandboxes optimized for native MCP execution.
Your AI agents in Claude Desktop only access the data you authorize, with DLP that blocks sensitive information from ever reaching the model, kill switch for instant shutdown, and up to 60% token savings. Enterprise-grade infrastructure, zero maintenance.
* Every MCP server runs on Vinkius-managed infrastructure inside AWS - a purpose-built runtime with per-request V8 isolates, Ed25519 signed audit chains, and sub-40ms cold starts optimized for native MCP execution. See our infrastructure
How Vinkius secures
Cohere for Claude Desktop
Every tool call from Claude Desktop to the Cohere MCP Server is protected by DLP redaction, cryptographic audit chains, V8 sandbox isolation, kill switch, and financial circuit breakers.
Frequently asked questions
How do I get a Cohere API Key?
Log in to the Cohere Dashboard, go to API Keys and click Create API Key. Copy the key immediately — it starts with a random string and won't be shown again. Free tier includes trial access with rate limits.
What models are available?
Use the list_models tool to see all available Cohere models. Key models include command-r-plus (most capable, 128K context), command-r (efficient, 128K context), command-r7b (lightweight, 128K context), embed-v4 (embeddings) and rerank-v3.5 (reranking).
Can I send multi-turn conversations?
Yes! Pass a messages array with alternating 'user', 'assistant' and 'system' roles. Each message has a 'role' and 'content' field. Command models support function calling and will return tool_calls when appropriate.
What is reranking and when should I use it?
Reranking reorders a set of documents by their relevance to a query. Use it after an initial search to improve result quality. The rerank tool takes a query, list of documents and returns them ranked by relevance score. Cohere's rerank models are industry-leading for search applications.
How does Claude Desktop discover MCP tools?
When Claude Desktop starts, it reads the claude_desktop_config.json file and connects to each configured MCP server. It calls the tools/list endpoint to fetch the schema for every available tool, then surfaces them as clickable options in the chat interface via the 🔌 icon.
What happens if the MCP server is temporarily unavailable?
Claude Desktop handles disconnections gracefully. if the server is unreachable at startup, the tools simply won't appear. Once the server becomes available again, restarting Claude Desktop will re-establish the connection. There is no timeout penalty or error loop.
Can I connect multiple MCP servers simultaneously?
Yes. You can add as many servers as you need in the mcpServers section of the config file. Each server appears as a separate tool provider, and Claude can use tools from multiple servers in a single conversation turn.
Is there a limit on the number of tools per server?
Claude Desktop can handle hundreds of tools per server. However, for optimal LLM performance, Vinkius servers are designed to expose focused, well-documented tool sets rather than overwhelming the model with too many options.
Does Claude Desktop support Streamable HTTP transport?
Yes. Claude Desktop supports both SSE (Server-Sent Events) and the newer Streamable HTTP transport that Vinkius uses. Simply provide the server URL. Claude auto-negotiates the transport protocol.
Server not appearing after restart
Ensure the JSON is valid (no trailing commas). Check the file path: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\\Claude\\ (Windows).
Authentication error
Verify your Vinkius token is correct. Go to cloud.vinkius.com to regenerate it if needed.
Tools not showing in chat
Click the 🔌 icon at the bottom of the chat input. If it shows 0 tools, the server may still be connecting. wait a few seconds.