Support for Claude subscriptions (via Claude Code/CLI)
It would be great to have an option to use an Anthropic subscription for Claude models instead of requiring API access.
Right now we can already connect a Chatgpt subscription via Codex CLI and a Gemini subscription via Gemini CLI, so it feels like a natural next step to support Claude subscriptions via Claude Code/CLI as well. I’m curious why there isn’t a similar option to use a Claude subscription through Claude Code/CLI.
Honestly, this is the one missing feature that keeps BoltAI from being perfect for me, and it's the only thing holding me back from purchasing a Team Perpetual license for my department. It's also the main reason I've been tempted to switch to ChatWise, since it supports all three subscriptions (Claude, Chatgpt, and Gemini). But BoltAI is by far the better app, and I'd much rather stick with it.
Is there a technical limitation that prevents BoltAI from using Claude Code/CLI in the same way it uses Codex CLI and Gemini CLI? Since Chatwise appears to support Claude subscriptions, I assume this should be technically possible, right?
@Daniel Nguyen, is this something you would consider adding to BoltAI? I think it would be an extremely valuable feature, especially for users who already pay for Claude subscription. I would love to get your opinion on this.
Log in to comment and vote
Comments2
Anıl Yuvalı
Sep 1
Hi @Daniel Nguyen, thanks for the info. This is awesome news. I just tried enabling the flag you mentioned and it worked perfectly. Thanks so much for your help. 😊
Daniel Nguyen
Sep 1
Hey @Anıl Yuvalı, the Claude Code integration is behind a feature flag. You can enable it in Settings > Developer > Feature Flag.
Unlike Codex and other CLIs, Anthropic is very strict about using their Claude Code subscription. It seems to be better now but they used to ban people left and right due to this. So much that people call them “banthropic”.
They seem to stop doing it now so I probably will move it out of the feature flag. But please note that there is still some risk using it.
Please enable the flag and let me know if it works for you.