Skip to content

Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦

C

Merge System One (one decision API, /v1/brio removed) into GLiNER2.5-Decide

JustVuggdce7a9fa4997
diff --cc docs/api.md
index 1654215d,691a709f..5d00d856
--- a/docs/api.md
+++ b/docs/api.md
@@@ -22,11 -22,11 +22,12 @@@ curl http://127.0.0.1:8000/v1/chat/comp
```
Implemented endpoints are `GET /v1/models`, `GET /v1/models/{model}`,
- `POST /v1/chat/completions`, legacy `POST /v1/completions`, `POST /v1/brio`
- (closed-set scoring, [systemone.md](systemone.md)) and `POST /v1/systemone`, the
- request and reply of TypeSafe's Jev API served by the same channel, or answered
- natively by a decision model such as [Laya](laya.md) or
+ `POST /v1/chat/completions`, legacy `POST /v1/completions` and
+ `POST /v1/systemone`, colibri's one decision API ([systemone.md](systemone.md)):
+ the request and reply of TypeSafe's Jev API, scored by a language model through
+ its logprob channel or answered natively by a decision model such as
-[Laya](laya.md). Chat and
++[Laya](laya.md) or
+[GLiNER2.5-Decide](gliner_decide.md). Chat and
completion requests support JSON responses, SSE streaming, usage counts,
`max_tokens`/`max_completion_tokens`, `temperature`, `top_p`, and up to four
custom `stop` sequences. `max_tokens` is a ceiling, not a target: when the
diff --cc docs/systemone.md
index fff9dd1b,e393498f..224a5d40
--- a/docs/systemone.md
+++ b/docs/systemone.md
@@@ -510,10 -461,9 +461,10 @@@ cannot tell which kind answered
| engine | model | doc |
|---|---|---|
| `c/laya` | Laya (Convai Innovations): ModernBERT encoder + decision head | [laya.md](laya.md) |
+| `c/gliner_decide` | GLiNER2.5-Decide (fastino): DeBERTa-v3 encoder + GLiNER2's classification head | [gliner_decide.md](gliner_decide.md) |
- A decision engine serves `POST /v1/systemone` only. Chat, completions,
- messages and `/v1/brio` answer 400 with a pointer to `/v1/systemone`, and
+ A decision engine serves `POST /v1/systemone` only. Chat, completions and
+ messages answer 400 with a pointer to `/v1/systemone`, and
`/v1/models` lists the model with `capabilities: ["systemone", "decision"]`
(a language model says `["chat", "systemone"]`). The path is one round trip:
the gateway checks the request, sends it to the engine as one record, and