PrismSearch
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| BRAVE_API_KEY | No | Brave Search API key (comma-separated for rotation); Brave is skipped when unset | |
| GOOGLE_CSE_ID | No | Google Custom Search search engine ID (cx) | |
| GOOGLE_API_KEY | No | Google Custom Search API key (comma-separated for rotation) | |
| PRISMSEARCH_HOME | No | Data directory, default ~/.prismsearch; holds the default cache database cache.db and the user config config.* | |
| PRISMSEARCH_CONFIG | No | Config file path (YAML / JSON / TOML); the CLI positional argument takes precedence | |
| PRISMSEARCH_LOG_LEVEL | No | Log level: error | warn | info | debug | |
| PRISMSEARCH_AUTH_TOKENS | No | Comma-separated access tokens; a non-empty value turns on security.auth automatically. HTTP clients send Authorization: Bearer <token> | |
| PRISMSEARCH_SERVER_HOST | No | HTTP bind address (default 127.0.0.1) | |
| PRISMSEARCH_SERVER_PORT | No | HTTP port (default 3000) | |
| PRISMSEARCH_WORKER_TOKEN | No | Shared fetch-worker token; used when workers.endpoints[].token is empty. Workers read the same variable | |
| PRISMSEARCH_SERVER_TRANSPORT | No | Transport: stdio (default) or http |
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": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| web_searchA | Search the web across multiple engines (Google, Bing, Brave, Baidu, DuckDuckGo, Sogou, 360, Yandex) with automatic engine selection and fallback. Returns titles, URLs and snippets. |
| image_searchA | Search images across engines that support it. Optionally inline small thumbnails as base64 data URIs. |
| fetch_pageA | Fetch a public web page and return it as markdown, plain text or HTML. Private/internal addresses are blocked. Use method=browser for JavaScript-rendered pages. |
| wechat_searchA | Search WeChat Official Account (微信公众号) articles via Sogou. Returns titles, account names, dates and short-lived signed URLs — fetch them promptly with fetch_page. |
| list_enginesA | List all configured search engines with their enabled state, health, latency, success rate and supported tools. |
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 5 tools
Each tool has a clearly distinct target: general web search, image search, WeChat-article search, raw page fetching, and engine introspection. The two search variants (wechat_search vs web_search) are explicitly scoped by source, so misselection is unlikely. fetch_page and list_engines occupy non-overlapping roles.
All names are snake_case and readable, with a coherent modifier+action pattern (wechat_search, image_search, web_search). The only deviation is list_engines, which puts the verb first, but it still fits the overall convention.
Five tools is a tight, well-scoped surface for a multi-engine search server, with no redundant or filler entries. Every tool maps to a distinct capability an agent would need.
The surface covers web, image, and WeChat search plus page retrieval and engine health — the core lifecycle for a search server. Minor gaps exist around result pagination/limits, batch fetching, or non-WeChat site-specific search, but agents can work around these via parameters and repeated calls.