dfine-semantic
OfficialServer Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| SEMANTIC_DATA_DIR | No | Where the indexes and the model are stored. | ~/.dfine-semantic |
| SEMANTIC_ALLOWED_ROOTS | No | Extra absolute roots the server may index (comma-separated). The working directory counts as a root unless it is / or your home folder. | cwd, ~/.claude |
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 |
|---|---|
| semantic_searchA | Find code by meaning in an indexed project. Ask one full natural-language sentence, not keywords. Good: "How does the app validate share token permissions?" Bad: "shareToken auth validate". Returns file:line references by default - read the files you need. Searches .ts and .tsx unless include adds other indexed extensions. |
| index_projectA | Build or refresh the semantic index of a project root. Embeds only new, edited or outdated files and drops deleted ones. Pass duplicates: true once to enable find_duplicates for this project. |
| index_statusA | List indexed projects with file, chunk and size counts - pass path for one project. |
| find_duplicatesA | List code in other files that nearly matches the given files or line ranges. Use it in reviews and refactors - pass changed line ranges ("src/a.ts:10-40") for sharper pairs. Treat every pair as a candidate - read both ranges before you call it a duplicate. Expect it to be off per project - ask the user before you turn it on. Turn it on with index_project duplicates: true - the first build can take over an hour on large projects. Covers .ts, .tsx, .js, .jsx, .mjs files; skips tests, specs and .d.ts files. |
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 4 tools
Each tool has a clearly distinct purpose: semantic_search retrieves code, index_project builds/refreshes the index, index_status reports state, and find_duplicates detects near-matches. There is no overlap between searching, indexing, status, and duplicate detection.
All names use snake_case consistently, but the internal pattern varies slightly (adjective_noun, noun_verb, noun_noun, verb_noun). Still highly readable and predictable as a set.
Four tools is lean but well-matched to a focused semantic-search-and-indexing server. Each tool earns its place without redundancy, though the surface is on the thin side.
The core lifecycle (index build, status, search, duplicate detection) is covered. Minor gaps exist, such as removing or clearing an index and per-file reindexing, but agents can work around these via index_project.