Kimi K3 MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| K3MCP_MAX_TOKENS | No | Maximum completion tokens per call | 16000 |
| OPENROUTER_MODEL | No | Model slug | moonshotai/kimi-k3 |
| K3MCP_MAX_ATTEMPTS | No | Total attempts for retryable failures | 3 |
| OPENROUTER_API_KEY | Yes | OpenRouter credential | |
| OPENROUTER_APP_NAME | No | OpenRouter attribution title | k3mcp |
| OPENROUTER_BASE_URL | No | API root | https://openrouter.ai/api/v1 |
| OPENROUTER_SITE_URL | No | OpenRouter attribution referrer | repository URL |
| K3MCP_MAX_INPUT_CHARS | No | Combined system and user prompt limit | 400000 |
| K3MCP_TIMEOUT_SECONDS | No | Per-attempt HTTP timeout | 600 |
| K3MCP_REASONING_EFFORT | No | OpenRouter reasoning effort | max |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| review_algorithmA | Only when explicitly requested, ask Kimi K3 to challenge an algorithm. |
| review_codeA | Only when explicitly requested, ask Kimi K3 to review code or a diff. |
| plan_projectA | Only when explicitly requested, ask Kimi K3 to create a project plan. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
Each tool targets a distinct task: algorithmic review, code review, and project planning. There is no overlap or ambiguity between them.
All tools follow the same verb_noun pattern (review_algorithm, review_code, plan_project), making the set predictable and easy to understand.
With three tools, the server is well-scoped for its purpose of invoking specific Kimi K3 actions. Each tool serves a clear function, and the count is within the ideal range.
The set covers key developer workflows (review and planning), but it lacks additional common tasks like code generation or debugging, which are minor gaps for the apparent scope.