TypeSafe Jev is the hosted model that introduced System One models on September 15, 2026: you send a state and typed questions, and it returns answers with probabilities instead of generated text. It is in early access and billed per token (third-party write-ups cite $0.042 per million input tokens).
Within days, open-source Jev alternatives appeared that do the same job on your own hardware. This post lists the ones worth testing, which of them accept Jev API requests unchanged, and what it takes to self-host one.
Why look for a Jev alternative?
- Data stays in your infrastructure. Tickets, emails and user messages never leave your servers.
- Cost at volume. A self-hosted model costs the same whether you make ten decisions or ten million.
- No waitlist. Open models are downloadable today.
- Fine-tuning. You can train an open model on your own labelled decisions. Jev's weights are not public.
- Offline and edge use. Some models run on a laptop CPU or an Apple Silicon Mac.
Best Jev alternatives to try first
If you already call Jev, start with a model that serves the same POST /v1/systemone endpoint, so existing client code only needs a new base URL:
| Model | Size | License | Serves the Jev API | Good first choice when |
|---|---|---|---|---|
| Laya | 322M / 421M | Apache 2.0 | Yes, through laya-serve | You want a small model that runs on CPU, or 100+ languages |
| Kev | 0.8B to 27B | Apache 2.0 | Yes, the TypeSafe SDK works unchanged | You have a GPU and want the closest drop-in |
| Decider | 2B / 4B / 35B MoE | Apache 2.0 | Yes | You serve with vLLM and want many-option choices |
| Von | 395M | Apache 2.0 | Yes | You need low latency on modest hardware |
| OpenThai-SystemOne | 0.8B | Apache 2.0 | Yes | Your input is Thai |
Other open options use their own APIs: Bespoke Nimble (9B), GLiNER2.5-Decide (340M), Together AI's Tev1-4B, and the training-free AnyJev and SemIf. See the full System One model comparison for all twelve.
Jev-compatible API servers
These projects put a Jev-style HTTP API in front of an open model, so you can switch models later without touching client code:
| Server | Language | Notes |
|---|---|---|
| laya-serve | Python | Ships with the upstream Laya package. See Self-host Laya. |
| Arbiter | Python | Routing, batching, metrics and a built-in playground |
| sys1 | Rust | candle-based, token batching, CPU, CUDA and Metal |
| ollaya | Rust | Ollama-style CLI and daemon; pulls models by name |
| laya-server | TypeScript | Docker image with a web console and API keys |
For the two Rust servers side by side, see Run Laya in Rust.
How to switch from Jev to a self-hosted model
- Pick a model from the table above and start its server locally.
-
Point your client at it. With the TypeSafe SDK, change the base URL; ollaya, for example, documents
TYPESAFE_BASE_URL=http://localhost:11435. With plain HTTP, the request body stays the same:
curl -X POST http://localhost:8080/v1/systemone \
-H 'Content-Type: application/json' \
-d '{"state":{"message":"I was charged twice"},"questions":{"refund":{"type":"noul","instructions":"Does the customer ask for a refund?"}}}'
- Compare on your own data. Send the same labelled requests to Jev and to the new model, and compare accuracy and latency before moving production traffic.
- Recalibrate thresholds. Probabilities from different models are not interchangeable. If you automate a decision above a confidence cut-off, set that cut-off again on held-out data.
Jev vs open alternatives: what you give up
Hosted Jev still leads out of the box on several published comparisons, especially on tasks with many labels. On Banking77, with 77 intents, upstream Laya reports 0.425 at default settings against Jev's 0.870. Open models close much of the gap after fine-tuning on your own data, and they win on cost, privacy and latency when you run them close to your application. See Laya vs Jev for the details and caveats.
Benchmarks in this space are young and mostly self-reported on different datasets, so treat any single number as a starting point, not a verdict.
FAQ
Is there an open-source version of Jev?
No. Jev's weights are not public. The Jev alternatives above are independent open models, mostly Apache 2.0, that answer the same kind of typed questions. Several of them also accept the same POST /v1/systemone requests.
Which Jev alternative runs on a CPU?
The encoder models are the lightest: Laya (322M and 421M parameters), Von (395M) and GLiNER2.5-Decide (340M). See Laya on CPU for measured latency.
Can I keep using the TypeSafe SDK?
Yes, with a server that speaks the Jev API. Kev and ollaya say the SDK works unchanged, and laya-serve says an existing Jev client only needs a new base URL.
Related
- System One models compared: all twelve models side by side
- Laya vs Jev: open weights, speed, accuracy and tradeoffs
- Self-host Laya: run a Jev-compatible API yourself
- Laya on CPU: latency and sizing without a GPU
Last verified: September 25, 2026.
I maintain laya-ai.com, an independent Laya resource site; this article first appeared there as Jev alternatives. The site also keeps an auto-updated list of System One models on GitHub. It was drafted with AI assistance, and every fact was checked against the project pages linked above.
Top comments (0)