The Generator Update

Features

• Modular, generator-based foundation (rewrote entire codebase)
• Significantly easier to build Open Interpreter into your applications via interpreter.chat(message) (see JARVIS for example implementation)
• Run interpreter --config to configure interpreter to run with any settings by default (set your default language model, system message, etc)
• Run interpreter --conversations to resume conversations
• Budget manager (thank you LiteLLM!) via interpreter --max_budget 0.1 (sets max budget per session in USD)
• Change the system message, temperature, max_tokens, etc. from the command line
• Central /conversations folder for persistent memory
• New hosted language models (thank you LiteLLM!) like Claude, Google PaLM, Cohere, and more.

What's Changed

• Fix typo 'recieved'> 'received' by @merlinfrombelgium in #361
• Pull request template created by @TanmayDoesAI in #365
• docs: move pr template to .github folder by @jordanbtucker in #373
• chore: enhance .gitignore by @jordanbtucker in #374
• chore: add vscode debug support by @jordanbtucker in #375
• discard the / as command as it will block the Mac/Linux to load the file by @moming2k in #378
• Update interpreter.py for a typo error by @YUFEIFUT in #397
• Translated Open Interpreter README into Hindi by @zeelsheladiya in #417
• Add models to pull request template by @mak448a in #423
• Retry connecting to openai after hitting rate limit to fix #442 by @mathiasrw in #452
• Handle %load_message failure in interpreter.py by @richawo in #431
• add budget manager for api calls by @krrishdholakia in #316
• The Generator Update by @KillianLucas in #482

New Contributors

• @YUFEIFUT made their first contribution in #397
• @zeelsheladiya made their first contribution in #417
• @mak448a made their first contribution in #423
• @mathiasrw made their first contribution in #452
• @richawo made their first contribution in #431
• @krrishdholakia made their first contribution in #316

Full Changelog: v0.1.4...v0.1.5