terse

Shrinks your AI agent's tool output so it burns fewer tokens, without changing what the agent reads. Lossless by default, no decode step

A new way for AI to search for data

View on GitHub

2GitHub stars

378commits on main

1forks

v0.49.0latest release, 2026-10-30

Python 99.4%, Shell 0.4%, Dockerfile 0.1%, JavaScript 0.1% · created 2026-06-27 · last push 2026-10-30

What XERJ indexed

Snapshot

Whole repository

The whole repository at commit 069059f585 (2026-09-30). main is 56 commits ahead.

Selected work

Eric Minish
@inth3shadows

“withhold the primer from offloaded results and stop count…”. 3 files, 348 lines in tests, src/terse.

Licence

The repository

MIT

LICENSE file checked at the pinned commit; the author agreed to inclusion.

The XERJ corpus

MIT, with credit

The author agreed by email on 2026-10-25. Same terms as the repository. Answers name the repo, the commit and the file:line.

selected for the XERJ Corpus Hublicence clearedsuggested use: adapt with attribution

agentcompressionjsonllmlosslessmcpproxytokens

Example questions

How does terse implement agent?
What does src/terse do in terse?
Show the compression code in terse.

Inside the indexed corpus

Languages by size: Python 99.4%, Shell 0.4%, Dockerfile 0.1%, JavaScript 0.1%, HTML 0%.
On GitHub since 2026-06-27, last push 2026-10-30, 378 commits on main.
The pin is 56 commits behind main; the corpus answers from the pinned code, not the newest.

Licence on GitHub and in the corpus

GitHub reported MIT for inth3shadows/terse on 2026-10-29, and the LICENSE file at commit 069059f585 was reviewed as MIT. They match, so answers from this corpus carry the same terms as the live repository.

Contributors

Eric Minish (@inth3shadows) has 3 selected repositories on the hub.

Their other repository pages: inth3shadows/runecho, inth3shadows/codeshot.

Using XERJ with terse

Code search MCP for Claude Code: cite file:line, not grep terse is tagged mcp on GitHub.
What is an MCP memory server?
A local search tool for Codex and Cursor terse is tagged claude-code on GitHub.
Local code search MCP that uses fewer tokens than grep

The prompt

Paste it into your AI agent. It clones inth3shadows/terse at the pinned commit and indexes it on your machine.

Search the GitHub repository inth3shadows/terse with XERJ (https://hub.xerj.org/corpus/inth3shadows--terse).
Run these in a shell, in order, and check each result before the next:
1. xerj --help  (success: it prints usage). If xerj is missing: curl -fsSL https://xerj.org/get | sh
2. git clone https://github.com/inth3shadows/terse terse && git -C terse checkout 069059f585b4255f94d23ca005ef84bee1f9431c
3. Start a local node and leave it running: xerj --insecure --data-dir ./xerj-data
4. xerj autoindex ./terse --url http://127.0.0.1:9200 --prefix terse --progress plain
5. xerj autoindex map --url http://127.0.0.1:9200
Then answer my question from the indexed files and cite the ax_path of every hit. Query syntax: https://xerj.org/llms.txt
Licence: MIT. Keep the copyright notice and credit the authors.

Eric Minish AI, prompts and agents corpora camel cognita agentops headroomlabs-ai/headroom marcusaguiarmaciel-cmd/cato adelinamart/robrain

Stats: GitHub API, fetched 2026-10-29. Selected work and consent: the XERJ registry.