QT/sure
1
0
Fork 0
mirror of https://github.com/we-promise/sure.git synced 2026-07-27 12:12:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

Revert "Refactor application workflows and update test coverage"

Browse Source 
This reverts commit 565e049f89.
This commit is contained in:
Juan José Mata
2026-07-23 13:39:46 -07:00
parent 565e049f89
commit c6eb7cdeed
15 changed files with 902 additions and 1166 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
docs/hosting/ai.md
View File

@@ -48,22 +48,6 @@ The easiest way to get started with AI features in Sure is to use OpenAI:
That's it! Sure will use OpenAI's with a default model (currently `gpt-4.1`) for all AI operations.
## Alternative: Codex Subscription OAuth
Sure can send Responses API inference through a ChatGPT-backed Codex subscription instead of billing an OpenAI Platform API key:
```bash
OPENAI_OAUTH_TOKEN=your-codex-oauth-access-token
# Optional when the account ID cannot be read from the token:
OPENAI_ACCOUNT_ID=your-chatgpt-account-id
# Optional; defaults to the current Codex default:
OPENAI_MODEL=gpt-5.6
```
When `OPENAI_OAUTH_TOKEN` is present, it takes precedence over `OPENAI_ACCESS_TOKEN`. Sure sends requests to the Codex Responses endpoint and adds the ChatGPT account header required by subscription authentication. This credential is only used for LLM inference; it is not treated as a general OpenAI Platform API key for hosted vector stores or other Platform APIs.
OAuth access tokens expire. Sign in again with an official Codex client and replace the configured token when that happens. Store the token like a password and never commit it to an environment file.
## Local vs. Cloud Inference
### Cloud Inference (Recommended for Most Users)