{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreiaiqxa26vcewbeuk3jvwudab6253ua4gw57nve4w5odf5jj7h3upe",
"uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mj3vtspdudi2"
},
"path": "/t/feature-request-let-codex-build-and-embed-small-task-specific-local-models/1378578#post_2",
"publishedAt": "2026-04-09T22:01:17.000Z",
"site": "https://community.openai.com",
"tags": [
"@Curoc1"
],
"textContent": "Hey @Curoc1, love the local-first angle. You’re basically proposing a “bundle a lightweight offline model” option in Codex (prep/adapt/quantize + integrate into the project), which would be a big win for offline and privacy-heavy use cases.\n\nI don’t have a timeline to share, but feedback like this is solid signal for where things could improve, and I’ll pass it along internally.\n\n-Sunny",
"title": "Feature Request: Let Codex Build and Embed Small Task-Specific Local Models"
}