Terse compresses every prompt before it reaches Copilot — removing filler, fixing typos, and stripping redundancy. Less context in means faster responses and lower cost on Copilot Business and Enterprise plans.
Copilot's context window fills up fast: your file, selected code, conversation history, and your typed prompt. Terse compresses the prompt text before it's sent — reducing tokens without reducing intent.
Terse detects VS Code via the Accessibility API and activates when you open the Copilot Chat panel or inline chat. No configuration required — it works the moment you start typing.
AutomaticSpell correction, whitespace normalization, filler word removal, hedge stripping, phrase shortening, redundancy elimination, and cleanup — in sequence, on every prompt you type to Copilot.
130+ RulesTerse's Agent Monitor shows token usage per Copilot Chat turn in real time. Know immediately when a single question is consuming 60,000 tokens — and why.
Live MonitorTerse tracks when you ask Copilot the same question twice in a session, or ask about the same file or function multiple times. Each duplicate is flagged so you can reuse the previous response instead.
Session AwarenessWorks in both Copilot Chat and inline chat (@workspace queries, /fix, /explain). Long natural-language questions are trimmed without losing technical accuracy — code blocks are always protected.
Chat + InlineTerse reads your prompts via macOS Accessibility APIs and compresses locally. No text is sent to Terse's servers — your code stays on your machine, just like Copilot's own privacy guarantees require.
Zero CloudCopilot's plans are seat-priced, but token efficiency directly affects how much value you get per seat and how fast context windows fill during long sessions.
Everything about using Terse with GitHub Copilot.
Terse reduces token costs across every major AI coding tool.
30-day free trial. No credit card required until your trial ends. Cancel anytime.