cn-websearch-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| KIMI_MODEL | No | Model name for the Kimi provider. | |
| MIMO_MODEL | No | Model name for the MiMo provider. | |
| ZHIPU_MODEL | No | Model name for the Zhipu provider. | |
| KIMI_API_KEY | No | Your Kimi (Moonshot) API key. At least one provider API key is required to perform searches. | |
| KIMI_ENABLED | No | Enable or disable the Kimi provider. Accepts true/false/1/0/yes/no/on/off. | |
| MIMO_API_KEY | No | Your Xiaomi MiMo API key. At least one provider API key is required to perform searches. | |
| MIMO_ENABLED | No | Enable or disable the MiMo provider. Accepts true/false/1/0/yes/no/on/off. | |
| KIMI_BASE_URL | No | Base URL for the Kimi provider. | |
| KIMI_PRIORITY | No | Priority for the Kimi provider (higher numbers go earlier). | |
| MIMO_BASE_URL | No | Base URL for the MiMo provider. | |
| MIMO_PRIORITY | No | Priority for the MiMo provider (higher numbers go earlier). | |
| STEPFUN_MODEL | No | Model name for the StepFun provider. | |
| ZHIPU_API_KEY | No | Your Zhipu GLM API key. At least one provider API key is required to perform searches. | |
| ZHIPU_ENABLED | No | Enable or disable the Zhipu provider. Accepts true/false/1/0/yes/no/on/off. | |
| ZHIPU_BASE_URL | No | Base URL for the Zhipu provider. | |
| ZHIPU_PRIORITY | No | Priority for the Zhipu provider (higher numbers go earlier). | |
| KIMI_TIMEOUT_MS | No | Per-provider timeout override for Kimi. | |
| MIMO_TIMEOUT_MS | No | Per-provider timeout override for MiMo. | |
| STEPFUN_API_KEY | No | Your StepFun API key. At least one provider API key is required to perform searches. | |
| STEPFUN_ENABLED | No | Enable or disable the StepFun provider. Accepts true/false/1/0/yes/no/on/off. | |
| WEBSEARCH_COUNT | No | Default result count when a tool call omits 'count'. Defaults to 8. | |
| WEBSEARCH_ORDER | No | Comma-separated provider priority order, highest first, e.g. 'stepfun,zhipu'. If not set, fallback to provider 'priority' values or alphabetical order. | |
| STEPFUN_BASE_URL | No | Base URL for the StepFun provider. | |
| STEPFUN_PRIORITY | No | Priority for the StepFun provider (higher numbers go earlier). | |
| WEBSEARCH_CONFIG | No | Path to an explicit JSON config file. If not set, cn-websearch.config.json in the working directory is auto-detected. | |
| WEBSEARCH_DEDUPE | No | Whether to merge duplicate URLs when aggregating. Accepts true/false/1/0/yes/no/on/off. Defaults to true. | |
| ZHIPU_TIMEOUT_MS | No | Per-provider timeout override for Zhipu. | |
| STEPFUN_TIMEOUT_MS | No | Per-provider timeout override for StepFun. | |
| WEBSEARCH_STRATEGY | No | Search strategy: 'fallback' (default, first success wins) or 'aggregate' (multi-source merge). | |
| ZHIPU_SEARCH_ENGINE | No | Zhipu search engine to use. Defaults to 'search_std'. | |
| WEBSEARCH_TIMEOUT_MS | No | Budget per provider attempt in milliseconds. Defaults to 30000. A retry gets a fresh budget. | |
| WEBSEARCH_MAX_PROVIDERS | No | Cap on providers per call (chain length / fan-out). Defaults to 4. |
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 | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| web_searchA | Search the web through multiple Chinese LLM providers (any of: Kimi, MiMo, StepFun, Zhipu GLM), depending on which ones you configured. Two strategies: 'fallback' tries providers in your configured priority order and returns the first success; 'aggregate' queries several providers in parallel and merges the results (deduplicated by URL, each item tagged with its source provider). Per-attempt timeout, one retry on transient failures. Returns normalized results { title, url, snippet, content?, published_date?, source? } plus _meta with the answering provider(s), total latency and a per-attempt audit trail. |
| provider_statusA | Read-only status: effective strategy and settings, plus which providers are enabled, have API keys and are part of the active search chain. |
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 2 tools
The two tools are clearly distinct: web_search performs searches, while provider_status reads configuration and provider health. There is no overlap or ambiguity between them.
Both names are concise, readable, and use snake_case consistently. web_search follows a verb_noun pattern while provider_status is noun-based, but they remain predictable and clearly tied to the server's purpose.
With only two tools, the server sits at the low end of acceptable scope. It feels slightly thin, but the narrow search-focused purpose makes the minimal count defensible.
The domain is a stateless web search service, and the two tools fully cover its needs: performing searches and checking provider status/configuration. There are no obvious dead ends or missing lifecycle operations.