Schema | Squeze

Squeze

Describes the environment variables required to run the server.

Name	Required	Description	Default
`root`	Yes	Absolute path to the repository to index and serve. Passed to the MCP server as `--root /absolute/path/to/repository`.
`tools`	No	Optional tool set selection. Use "all" to expose the extra `squeze_compact` and `squeze_json` tools in addition to the default four tools.	default
`budget`	No	Token budget for context packs, 256–32,000. Default 3,000.	3000
`features`	No	Comma-separated feature list, e.g. `-prf,-priors`, `+frames,+git`, or `none` (closest to 2.1.0 behaviour).
`RSQ_FEATURES`	No	Comma-separated feature list for ranking and packing, e.g. `-prf,-priors`, `+frames,+git`, or `none`.
`RSQ_TOKENIZER`	No	Set to `v1` to use the 2.1.0 tokenizer for comparisons.
`semantic-model`	No	Enable optional local Ollama embeddings for the 300 strongest lexical candidates. Supported value: `all-minilm`.
`REPOSQUEEZE_CACHE_DIR`	No	Cache location. Use `repo` to restore the old in-repo `.reposqueeze-v2/` cache directory.
`REPOSQUEEZE_EXACT_STATS`	No	Set to `1` to use the exact slow input-token statistic instead of the sampled estimate.
`REPOSQUEEZE_REFRESH_INDEX`	No	Set to `1` to force re-reading files in repositories above 5,000 files.
`REPOSQUEEZE_MAX_CANDIDATES`	No	Number of candidates parsed in large-repository mode, 100–5,000.

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.

Server capabilities have not been inspected yet.

Functions exposed to the LLM to take actions

Name	Description
No tools

Interactive templates invoked by user choice

Name	Description
No prompts

Contextual data attached and managed by the client

Name	Description
No resources

This server cannot be deployed

ActivityMaintained

ResponsivenessNo issues