GLiClass vs kettei

GLiClass

An Apache-2.0 classifier that scores a whole label set in one encoder pass on the owner's hardware, with single-label, multi-label, hierarchical and few-shot modes. It returns label scores with no calibration claim, the bundled server has no authentication, and the last three test runs on the main branch, on 24 September 2026, failed.

kettei

OneHot returns probabilities for up to 128 typed questions a call at $0.02 to $0.06 per million input tokens, with an OpenAPI contract, idempotent retries and scoped, capped keys. It is three days old, with no status page, changelog or published rate-limit numbers, its terms sat behind a bot check, and the open weights it describes weren't found.

GLiClass

1. Pass --host 127.0.0.1 to python -m gliclass.serve, or put the port behind your own gateway. The server checks no credential
2. On a machine without a GPU add --device cpu --dtype float32 --num-gpus-per-replica 0. The default configuration expects CUDA
3. Send one text a request to POST /gliclass. An array in texts is cut to its first item without an error
4. Set multi_label to false for one label from a set. The default scores each label independently, so scores do not sum to 1
5. Keep text plus labels under the pipeline's 1,024-token max_length, or use ZeroShotClassificationWithChunkingPipeline. Longer input is truncated silently

kettei

1. Send context and decisions to POST https://api.sqwish.ai/v1/decide with a bearer key. Leaving out model gets kettei-core, which reads text only
2. Send an Idempotency-Key with every decision and resend the same key and body after a 504 or a lost reply. Replays aren't charged
3. Branch on error.code and error.retryable, not the status. Wait for Retry-After on 429 and retryable 503s
4. Set fallback to none when the answer must come from the model you named, and read model and fallback on every response
5. Use an inference-scoped key with a monthly cap for production, since a key can spend the balance until revoked