The human programming language
Human is the first open source AI programming language that makes it possible to code using LLMs.
Setup
Requires Claude Code and uv.
uv tool install humanlang
human skills # installs the /human and /decompile skills into ~/.claude/skills
Upgrade to the latest version with:
uv tool upgrade humanlang
human skills # refresh the installed skills too
human init # run in each project to refresh its reader
Use
In your project:
human init # creates the human/ folder: the maps and the reader live there
Then in Claude Code:
/human <abstraction>— compile free text into code, mapped to your words./decompile <file>— explain an existing file, pinned to its lines.
Read
human serve
Open http://localhost:8010/human/web.html. Every entry opens as a notepad, and "compile" hands your words to claude. A "new file" line at the top of the file tree makes an empty file, in a new folder too; write its abstraction and claude writes the code under it.
Train
human train --open # start a session; /decompile now writes three versions per file
human train --close # map the versions you picked and finish the session
Pick in the reader at http://localhost:8010/human/web.html: an open session shows as a layer over it, swipe sideways for the versions, down for the next file, and say why you picked if you want. A "close training" button applies the picks; a new abstraction you do not pick takes best, a sync you do not pick carries over. A write in the reader that asks claude for a telling opens a session by itself. The sessions live in the training store, one JSON per session under your user, with a copy of the code and of the abstraction each row worked on; human login <key> once per machine names you.
Everything the project writes lives in the human/ folder — one rm -rf human/ removes it completely.
Release files for humanlang 0.0.33
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| humanlang-0.0.33.tar.gz | 144.9 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| humanlang-0.0.33-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 298.3 kB
Release files / humanlang-0.0.33.tar.gz
| Download URL | humanlang-0.0.33.tar.gz |
|---|---|
| Size | 144.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
ca6a5a8ef1c1a64bf899bfa7d938a177d3cd4d0db60c4154ea6a2e66957aeb8e
|
|
BLAKE2b-256 checksum How to use checksums |
f6dd09cf306b8ebca60d914c005ed0fc4cc4626e1658c9f53220abc0aa31db33
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.
Transparency logRelease files / humanlang-0.0.33-py3-none-any.whl
| Download URL | humanlang-0.0.33-py3-none-any.whl |
|---|---|
| Size | 153.4 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
ae02d71a6e46b2652045c71eb004708f136ba4cdf39b0b02773a19584e54e661
|
|
BLAKE2b-256 checksum How to use checksums |
a0e223aee94c622be5f32098f051553f3e2b6333d9c59a1fefbff3b2d15a2e1f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.
Transparency log