Releases: waylaidwanderer/node-chatgpt-api
Releases · waylaidwanderer/node-chatgpt-api
v1.15.0
This release adds support for Bing client in the CLI and fixes some bugs related to Bing ChatHub.
Features
- Add Bing client support in the CLI. See the updated settings example file.
Bug Fixes
- Handle errors from ChatHub gracefully. If ChatHub returns an error, the CLI will display this error.
- Check if Bing client is authorized before sending messages. If Bing client is not authorized, the CLI will display this error.
PS: Bing helped me write this! :D I talked to it in the CLI app and asked it to write me release notes based on the git log messages I gave to it.
Full Changelog: v1.14.0...v1.15.0
v1.14.0 - Reverse Proxy Support
Features
- add support for using a reverse proxy/private API server, which allows using ChatGPT's raw models directly again (see Using a Reverse Proxy)
Full Changelog: v1.13.0...v1.14.0
v1.13.0
Features
- I FOUND THE NEW MODEL NAME
- build(deps): bump fastify from 4.12.0 to 4.13.0 by @dependabot in #38
Full Changelog: v1.12.0...v1.13.0
v1.12.0
v1.11.0 - Bing's GPT-4 ChatGPT Client 🚀
Features
- Added experimental Bing client in #35 (see readme for more info)
Full Changelog: v1.10.1...v1.11.0
v1.10.1
Fixes
- handle SSE errors properly to avoid retry loop leading to 429s
Full Changelog: v1.10.0...v1.10.1
v1.10.0
Features
- support streamed responses (#25) (client, API, CLI)
- Readme updated with examples
- API: allow CORS all origins, closes #24
Fixes
- API: handle empty message properly
- CLI: try-catch boxen so malformed input doesn't crash the app
Full Changelog: v1.9.0...v1.10.0
v1.9.0
v1.8.0
Features
- Adds Docker support for the API server (thanks to @queercat for helping me get started)
- you can run it with
docker-compose up
. For available ENV vars, check the settings example file, or you can just set options in settings.js as usual.
- you can run it with
Full Changelog: v1.7.0...v1.8.0
v1.7.0
Features
- Make the AI obey prompt instructions better by always inserting
promptPrefix
before the last user message.- This is especially important for custom instructions. After a bunch of testing, it doesn't seem to cause the AI any confusion, even if you ask it things like "what's the last thing I wrote?".
Fixes
- Reduce the temperature to 0.8, producing more coherent results with longer outputs while allowing for decent creativity.
Full Changelog: v1.6.1...v1.7.0