Add Terse to Usage & Observability - #625
Open
lucaszengool wants to merge 1 commit into
Open
Conversation
Author
|
merged |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds Terse to
📊 Usage & Observability, placed by star count (29 ⭐, between Claude-Monitor and cccost).Why it's distinct from what's already listed: everything currently in this section reports usage after the fact. Terse also acts on it — a burn-rate/token/dollar ceiling escalates from an alert to
SIGSTOP-pausing orSIGTERM-killing the agent process before the next API call fires, so a loop stops at the ceiling rather than at the invoice. It also compresses the prompt before it's sent, and risk-scores MCP servers (remote transport, embedded credentials, code-execution surface) alongside their per-call token cost.Reads Claude Code's session JSONL for exact token counts, cache read/write efficiency, burn rate and context fill; also covers Cursor, Codex, Copilot CLI, Cline, Windsurf, OpenClaw and Aider. Runs fully on-device.
Disclosure: I'm the author. The desktop app is commercial (free 30-day trial); the SDK and the benchmark harness in the repo are MIT licensed and free —
npm run benchmarkreproduces the compression numbers locally.Happy to adjust the wording or drop it lower in the section if you'd prefer.