whereismyai
Jiwo
Jiwo is a software library providing decision models that read state and questions to return probability scores in one forward pass.
Jiwo offers non-generative decision models designed for latency-critical environments. It functions by accepting a state and a series of questions, returning probabilities for specific options, scores for ordered levels, or true/false confirmations. The system implements the Jev API format. Users can deploy it as a standalone HTTP server or integrate it directly into processes. It supports Qwen3.5-based models and accommodates LoRA adapters.
Jiwo at a glance
- Best for: Software engineer
- Runs on: Linux, macOS, Web
- Pricing: Free
- Released: October 2026
- Category: AI agent tool
- Website: github.com/jiwidi/jiwo
Key features
- Supports choice, score, and noul question types
- Performs inference in a single forward pass
- Provides HTTP API compatible with TypeSafe Jev
- Allows local checkpoint and Hugging Face model loading
- Includes server for handling API requests
- Supports optional API key authentication
Use it for
- Running latency-critical decision tasks
- Classifying state data into probability options
- Evaluating true or false statements via noul questions
- Ranking states on ordered levels using score types
Alternatives to Jiwo
- NanoJev — It runs parallel decisions and outputs probability distributions without output-token decoding.
- Jev — Fast, cheap typed decisions that software can act on directly
- SemIf OpenJev — Extracts direct token probabilities without generating text or running JSON repair loops.
- Decido — Runs locally on CPU while beating published benchmark scores.
- Vev — Runs locally on NVIDIA hardware with a fast, Jev-compatible API.
- ej — Runs locally on CPU from a single 11 MB file without APIs or GPUs.
- Kev — Provides calibrated probabilities and matches the TypeSafe API.
- Sprix Sage Router — It evaluates continuation, collaboration, and handoff routes in one objective.
- Laya Mlx — Provides local MLX inference with no PyTorch or cloud API dependency.
- Cerebellum — Routes decisions between local models and cloud models while working offline.
- Beeuni — Agents retain context and improve skills over time through a controlled review process.
- OpenSquilla — Maximizes AI intelligence density while minimizing token costs through smart routing.
All Jiwo alternatives