Laconic – Compress what a coding agent carries, not what it says.
Status: research and evaluation package, v0.7.0. The reversible codec core, replay/gate harness, and human-study harness are released. Hook integration is the intended primary runtime surface and MCP is secondary, but both are blocked: K1 measures 8.53% net savings on the committed fixture corpus, below the 15% kill threshold. That corpus validates the gate machinery, not real-world deployment economics.
Laconic is a codec for the machine-to-machine traffic of coding agents: it re-encodes observations and actions at the tool boundary, keeps every elision recoverable, and renders prose for a human only on demand.
docs/grounding.mdis the authoritative statement of what Laconic is, what it deliberately is not, and how to detect strategy drift. Read it before proposing or reviewing changes.
Install
uv tool install laconic
laconic --help
Documentation
| Document | What's in it |
|---|---|
docs/pitch.md |
The short version: the problem, the measurement, what is and isn't claimed |
docs/overview.md |
What Laconic is, why, the evidence, prior work, and the pre-registered gates |
docs/system-design.md |
Architecture, components, data model, and the evaluation harness |
docs/observe-design.md |
Laconic Observe: the automatic, content-free measurement surface design |
docs/observe-cli.md |
laconic observe operator guide: install/remove/status/report |
Start with evidence
Run the current measurement and gate surfaces before considering deployment:
uv run laconic measure tests/corpus --expect tests/corpus/expected.json
uv run laconic gates --corpus tests/corpus --format json
The K1 fixture verdict blocks live hook/MCP integration. A representative-corpus K1 decision is the next product gate; it is not a reason to tune the codec or relax thresholds after observing the fixture result.
License
Apache License 2.0 — see LICENSE.
1 maintainer
Metadata
1 maintainer
Release files for laconic 0.8.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| laconic-0.8.0.tar.gz | 245.6 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| laconic-0.8.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 387.7 kB
Release files / laconic-0.8.0.tar.gz
| Download URL | laconic-0.8.0.tar.gz |
|---|---|
| Size | 245.6 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
2198711e64ba4d4ebe098609f9e1de7f453ab8dd83ca1d6f62e08378b87cf190
|
|
BLAKE2b-256 checksum How to use checksums |
6fc61250c005434abd320d04918664a8a8741b7b038cb5019f620df3c526e627
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.6.14
|
Release files / laconic-0.8.0-py3-none-any.whl
| Download URL | laconic-0.8.0-py3-none-any.whl |
|---|---|
| Size | 142.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
e02a6e88140a3647fae7b3e6841df40af0f7d32b26639998f63a3582b31b37c7
|
|
BLAKE2b-256 checksum How to use checksums |
fcec2cdf61d4b60ddbc4c46839f5fea6d9b5c286064973fdce28ae12a3cc3230
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.6.14
|
