TinySearch
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MCP_HOST | No | Host address for the MCP server when using streamable-http or sse transport. | 0.0.0.0 |
| SEARXNG_URL | No | SearXNG-compatible search endpoint used by TinySearch. In the Glama bundled setup, this points to the local SearXNG process inside the same container. | http://127.0.0.1:8080/search |
| MCP_TRANSPORT | No | Transport protocol for the MCP server. For Glama, keep this as stdio because mcp-proxy wraps the server. | stdio |
| OPENAI_API_KEY | No | API key for OpenAI-compatible embedding service. | |
| OPENAI_BASE_URL | No | Base URL for OpenAI-compatible embedding API. | |
| SEARXNG_SETTINGS_PATH | No | Path to the SearXNG settings file used by the bundled Glama startup script. | /etc/searxng/settings.yml |
| TINYSEARCH_MODELS_DIR | No | Directory to store downloaded ONNX embedding models. | /data/models |
| OPENAI_EMBEDDING_MODEL | No | Model name for OpenAI-compatible embeddings. | |
| TINYSEARCH_CONFIG_PATH | No | Path to a custom research config JSON file. | |
| TINYSEARCH_ONNX_MODEL_DIR | No | Exact directory path for a specific ONNX model bundle. | |
| SEARXNG_STARTUP_WAIT_SECONDS | No | Number of seconds to wait for bundled SearXNG to start before launching TinySearch. | 5 |
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 |
|---|---|
| researchA | Search the web, crawl ranked pages, and return a grounded answer prompt. Input schema has exactly one field: query. Pass the user's question as-is. |
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 1 tool
Only one tool exists, so there is no risk of confusion. The tool's purpose is clear and unambiguous.
With a single tool, there is no inconsistency. The name 'research' is a reasonable verb that describes the action.
One tool for a tiny search server is slightly minimal but appropriate for the intended simplicity. The tool performs the entire search and answer generation pipeline.
The single tool covers the full workflow of searching, crawling, and returning a grounded answer, which is complete for the server's stated purpose. No obvious gaps.