feat(models): persona models same on chat + Telegram, no_think, one resident - #362
Merged
Merged
Conversation
…o_think Goal: one model resident at a time, switched in CODEC, and usable on Telegram — the same for every model — instead of a separate always-on server per persona. - codec_models.model_extras: add a `no_think` flag (extra_models[].no_think). - routes/chat.py: force thinking off for a no_think model, so a reformat/persona fine-tune answers directly instead of burning the budget on a <think> block. - codec_bridges.call_llm: when the active model is a persona model, the bridge (Telegram, etc.) speaks AS that persona and applies its sampling — mirroring the dashboard chat. So "switch the model in CODEC, then use it on Telegram" behaves identically. The everyday/operator model is unchanged (channel persona). M Corpus 4B + 8B and the 8B reformat voice are declared no_think in config (local). No new standalone server / port — a persona is served by the one switched model server, keeping a single model resident. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Makes every model behave the same: one resident at a time, switched in CODEC, usable on Telegram — no per-persona always-on server.
model_extras: newno_thinkflag.<think>).codec_bridges.call_llm: active persona model → bridge speaks AS it + applies its sampling (Telegram matches dashboard chat). Everyday model unchanged.M Corpus 4B + 8B + 8B reformat = no_think (config, local). No new port/server.
Verified: 8B reformat one-shot with /no_think → clean "Please forward the invoice… Mickaël". Tests: 61 passed.
Needs
pm2 restart codec-dashboard codec-telegram.🤖 Generated with Claude Code