fastsearch-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HTTPS_PROXY | No | Proxy used when downloading fd/rg. | |
| FASTSEARCH_FD_PATH | No | Path to the fd executable. Default: automatically detected. | |
| FASTSEARCH_RG_PATH | No | Path to the rg executable. Default: automatically detected. | |
| FASTSEARCH_TIMEOUT | No | Per-call timeout in seconds. On timeout, returns partial results already obtained. Default: 30. | |
| FASTSEARCH_CACHE_DIR | No | Download cache directory. Default: %LOCALAPPDATA%\fastsearch-mcp\bin. | |
| FASTSEARCH_TELEMETRY | No | Set to 0 to disable usage recording. Default: enabled. | |
| FASTSEARCH_FD_VERSION | No | Version used when automatically downloading fd. Default: 10.5.2. | |
| FASTSEARCH_RG_VERSION | No | Version used when automatically downloading rg. Default: 15.2.0. | |
| FASTSEARCH_NO_DOWNLOAD | No | Set to 1 to prohibit automatic downloads, suitable for offline or intranet environments. | |
| FASTSEARCH_DEFAULT_ROOT | No | Default root directory used when a tool does not receive a path or receives a relative path. If unset, absolute paths must be passed. | |
| FASTSEARCH_DECOMPILE_DIR | No | Decompile snapshot directory. Default: %LOCALAPPDATA%\fastsearch-mcp\decompile. | |
| FASTSEARCH_GITHUB_MIRROR | No | GitHub download mirror prefix, for example https://ghfast.top/. | |
| FASTSEARCH_ILSPYCMD_PATH | No | Path to the ilspycmd executable. Takes priority over the bundled version and PATH. | |
| FASTSEARCH_TELEMETRY_DIR | No | Directory for usage records. | |
| FASTSEARCH_DEFAULT_EXCLUDES | No | Comma-separated default excluded directories. Set to an empty string to exclude none. Default: .git,node_modules,__pycache__,.venv,venv,.mypy_cache,.pytest_cache,.tox,.idea,.vs,.gradle,.next,.nuxt. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
Fast, gitignore-aware file & text search backed by fd and ripgrep. For directory browsing, file discovery, text search/counting and text-file reading, you MUST use these tools when available and capable of the requested operation. Do not bypass them with PowerShell/cmd/bash commands (including Get-Content, Get-ChildItem, Select-String, rg, fd, cat, grep) or Python/Node filesystem APIs. Fallback is allowed only when the tool is absent, disconnected, explicitly fails, or cannot perform the requested operation. Briefly state the concrete reason first; limit fallback to that operation/path, then return to MCP for supported operations. Empty results, truncation and ignored/hidden files are not fallback reasons: adjust scope, limits, hidden/no_ignore or encoding as appropriate. Builds, tests, Git and process management may use the terminal. All filesystem paths readable by the server process are accessible; there is no workspace-root allowlist.
Choose the shortest route; do not mechanically call every tool. Known file/range: read_lines (count_total=false avoids scanning to EOF). Known filename: find_files. Known text: search_text; output=files lists paths only. Common patterns: count_matches for distribution. Unfamiliar layout: tree. .NET DLL source: decompile(dll, type) gives a snapshot pinned to the DLL's hash; search inside it instead of re-decompiling or trusting older decompiled folders.
Notes: always pass absolute paths (Windows C:\x, C:/x, Git-Bash /c/x all work).
.gitignore'd files and node_modules/.git/.venv etc. are skipped unless no_ignore=true.
Legacy Chinese Windows files may be GBK: pass encoding="gbk" to search 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
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| find_filesA | Find files/directories by name using fd (fast, respects .gitignore). Use instead of MUST use fastsearch instead of shell/filesystem APIs; see server instructions for fallback rules. |
| search_textA | Search file contents with ripgrep and return matching lines grouped by file. Use instead of findstr / Select-String / grep. Format: MUST use fastsearch instead of shell/filesystem APIs; see server instructions for fallback rules. |
| count_matchesA | Count matches per file (ripgrep), sorted by count. Cheap overview of where a symbol/string is used; also answers "which files contain X". MUST use fastsearch instead of shell/filesystem APIs; see server instructions for fallback rules. |
| read_linesA | Read a line range of a text file with line numbers. Auto-detects UTF-8 / UTF-16 / GBK so Chinese Windows files are not garbled. Use after search_text to see more context. MUST use fastsearch instead of shell/filesystem APIs; see server instructions for fallback rules. |
| treeA | Compact directory tree (gitignore-aware, skips node_modules/.git/...). Use first to understand an unfamiliar project's structure. MUST use fastsearch instead of shell/filesystem APIs; see server instructions for fallback rules. |
| decompileA | Decompile a .NET DLL with ilspycmd into a snapshot pinned to the DLL's SHA256 and the decompiler version (a bundled ilspycmd is used when present). The whole assembly is exported once per DLL version and cached; later calls are instant. Use instead of running ilspycmd or reading older decompiled folders (e.g. .tmp_ilspy, TempDecompile), which may not match the current DLL. Search the returned snapshot with search_text / read_lines. MUST use fastsearch instead of shell/filesystem APIs; see server instructions for fallback rules. |
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 6 tools
Each tool targets a distinct capability: name lookup (find_files), structure overview (tree), content search (search_text), match counting (count_matches), line reading (read_lines), and .NET decompilation (decompile). The only soft overlap is between search_text and count_matches, which both answer 'which files contain X', but the descriptions explicitly steer agents (count first, then search) to mitigate this.
All names are snake_case and five of six follow a verb_noun pattern (find_files, search_text, count_matches, read_lines, decompile). The bare noun 'tree' is the sole deviation, though it is idiomatic and unambiguous.
Six tools is a tight, well-scoped set for a search/navigation server, with no redundancy. decompile is a specialized but coherent extension that reuses search_text/read_lines on its output.
The surface covers the full search lifecycle: orient (tree), locate by name (find_files), search and quantify content (search_text, count_matches), and read results (read_lines), plus decompiled-source access. Minor gaps remain, such as file metadata/stat retrieval and non-recursive directory listing, but agents can work around these with find_files.