lcc (Local Context Compiler)
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TYPESAFE_API_KEY | No | API key for the remote TypeSafe API, needed only when using the opt-in `jev` provider (`--provider jev` / `provider: jev`). Not required for the default local providers. | |
| LCC_DISABLE_NETWORK | No | Set to 1 to disable network access (and therefore the optional `jev` provider). | 0 |
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 | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| compactB | Drop context blocks irrelevant to an objective. Default provider mechanical is fully offline (no key, no network); laya is offline semantic (needs the [laya] extra, else honest mechanical fallback); jev needs TYPESAFE_API_KEY. Fail-safe: failures keep content. |
| compact_transcriptA | Tool-call compaction for a session transcript: pairs each tool call with its result, pins the first and newest messages, and drops the pairs a semantic judge says are spent — kept messages stay verbatim, so nothing is summarized. Default 'auto' is the local specialist-v3 ranker (offline, no key); 'jev' is remote and needs TYPESAFE_API_KEY. Every failure keeps the call. |
| inspectA | Read-only diagnostic inspection of a text: tokens, structure, duplication, cleanup projection, and recommendation. Offline. |
| prepareC | Inspect-first deterministic prompt preparation: runs inspection, then lexical selection + safe optimization only when recommended. Never summarizes or paraphrases. Offline. |
| explainB | Explain a relevance-compaction report (as produced by compact): why every block was kept, trimmed, or dropped. Never re-runs, never touches the network. |
| recallA | Return a tool output that compact_transcript (rules_mode 'stashed') replaced with '[removed: N tokens · lcc recall ]'. The bytes are re-hashed on read, so the text is exactly what was removed. Offline, local stash only. |
| intakeC | Intake triage for raw prompt input: readiness, intent, clarifying questions, assumptions, and a compiled prompt. Offline by default. |
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 7 tools
Each tool has a distinct role in a coherent pipeline (compact for context blocks, compact_transcript for session transcripts, inspect/prepare/explain for diagnostics, recall for stash retrieval, intake for triage). The main risk is compact vs compact_transcript and inspect vs prepare, which overlap conceptually, but the descriptions clearly delineate scope and inputs.
All tools use a consistent lowercase verb pattern (compact, inspect, prepare, explain, recall, intake), with snake_case only for the compound compact_transcript. No mixed conventions or vague naming.
Seven tools is well-scoped for a context compiler, and each tool maps to a distinct stage (compaction, diagnostics, retrieval, intake) with no redundant entries.
The surface covers the core lifecycle: intake, inspection, preparation, compaction, explanation, and retrieval of stashed content. Minor gaps exist (e.g. no explicit way to list or enumerate stashed items before recalling, or a config/status tool), but core workflows are covered.