Skip to main content

lmtk

Interactively play with GPT-based tools from the terminal

PyPi License

Description

(TODO: Update README with information about other modes, config profiles, scripting, JS sandbox, and Jupyter integration)

ChatGPT needs a power-user mode. lmtk gives you a terminal UI for interacting with ChatGPT and other GPT-based tools.

The default mode, synth-chat, uses a ChatGPT-like chatbot built directly from GPT-3. It supports features not possible with ChatGPT, including:

  • Writing the prefix of the next response
  • Giving instructions that will always be followed and never forgetten

This project is still in the early stages of development. It will have bugs and frequent breaking changes.

Install

For now, Python >=3.9 is required. Use pip3 instead of pip if necessary.

pip install -U lmtk[extras]

If you are having trouble with the TensorFlow dependency, you can exclude it:

pip install -U lmtk

This will disable automatic code syntax detection, but most lmtk modes are good about manual syntax annotation.

If you don't have an OpenAI API key create one here and set it:

export OPENAI_API_KEY="<your api key>"

Usage

lmtk @thread-name [-m mode-name]

Use lmtk modes to list available REPL modes and lmtk threads to list open threads.

Inside the REPL, type .help for a list of commands and keyboard shortcuts.

Tips

Conversation Seeds

Conversation seeds give you extremely strong influence over the chatbot's behavior.

.seed You must write your message using only lowercase letters

All future responses will only use lowercase letters (usually). Seeds are never forgotten. Seeds are best phrased as statements of fact or commands. See .help for examples.

Forced Response Prefix

If you aren't getting a response you like, you can directly specify how the next response must start using :>. For example, if you send:

Give me Pong in Pygame :> Here is the code:

Then the response will start with Here is the code:.

Useful Commands

See .help for a full list, but these are some particularly important commands:

  • .clear or C-x + C-c to clear the screen
  • .exit or C-d to exit the REPL
  • .new or C-x + C-n to reset the thread
  • .redo or C-x + C-r to resample the response
  • .undo or C-x + C-u to rewrite your most recent message

Custom Modes

Add a Python file that looks like this to $LMTK_CONFIG_PATH/plugins/. By default this will be ~/.config/lmtk/plugins/.

See lmtk/modes for more complete examples.

Development

To install lmtk from source:

git clone git@github.com:veered/lmtk.git
cd lmtk
pip install -U flit
flit install -s

Screenshot

Screen Shot 2022-12-16 at 6 09 31 PM

Metadata

Release files for lmtk 0.4.32

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for lmtk 0.4.32
File Size Uploaded Provenance
lmtk-0.4.32.tar.gz 50.3 kB Without provenance Details

Built distribution (wheel)

Table of built distributions (wheels) for lmtk 0.4.32
File Interpreter ABI Platform
lmtk-0.4.32-py3-none-any.whl Python 3 none any Details

Total release size: 113.2 kB

Release files / lmtk-0.4.32.tar.gz

Download URL lmtk-0.4.32.tar.gz
Size 50.3 kB
Tags Source
SHA-256 checksum
How to use checksums
904a3e4fd352703fb161510883ae65fcd3dc556c07307bfc36b4b1efa2e6d042
BLAKE2b-256 checksum
How to use checksums
89d6a909c9ea04190f2ab15e7214f17454130c4edd97d3eee9a13345954ba684
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via python-requests/2.28.1

Release files / lmtk-0.4.32-py3-none-any.whl

Download URL lmtk-0.4.32-py3-none-any.whl
Size 62.9 kB
Tags Python 3
SHA-256 checksum
How to use checksums
b25a0194d59e0898e553bdeb524b433ebbc0526b245affdfb1a724f10abff7c9
BLAKE2b-256 checksum
How to use checksums
f009555242fc6ddf9a99b3e1fc9d9bbde038b0a0145df64505d9d09476113ca1
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via python-requests/2.28.1

Security for lmtk 0.4.32

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.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page