Updated

Flide is an open-source inference engine for Jev-style decision models.

Pricing Free open source
Visit site

Follow Flide to get an email when its listing changes, such as new pricing or features. One digest a day at most, and you can unsubscribe anytime.

Company OmniJev
Pricing model Open Source
Paid plans None listed
Category peers 9
01Overview

It serves the Jev `/v1/systemone` API and uses exact state forking to read a shared state once and answer many questions on it in one pass, returning a probability for every option so agents and pipelines can make thresholded decisions.

02Key features
  1. 01

    `pip install flide` and `flide serve --model ...` start an API server

  2. 02

    Jev `/v1/systemone` API with choice and yes/no questions and per-option probabilities

  3. 03

    Exact state forking so each answer matches the model asked alone

  4. 04

    Serves Qwen3.5 and Gemma 4 chat models plus OneJev, Open-Jev, pplx-decider, Clef and other decision models

  5. 05

    LoRA adapters and image inputs for supported families

  6. 06

    `flide verify` checks exact forking on your own model and requests

03Pricing

Open Source

Flide is free and open-source under the Apache-2.0 license and installs from PyPI. You run it on your own NVIDIA GPU; prebuilt kernels cover A10, A40, RTX 30, RTX 50, RTX PRO 6000, H100 and H200.

04Who it's for

Best for

Teams running Jev-style decision models for agent routing, triage, classification or LLM-as-a-judge workloads who need high throughput on their own GPUs.

Not ideal for

Developers looking for a coding assistant, general chat or text generation serving, or anyone without an NVIDIA GPU.

05Verdict
A fast-moving, specialised serving engine for the Jev decision model ecosystem; its speed claims are vendor benchmarks and it is still pre-1.0.
ai.dosa.dev editorial · Nov 10, 2026
06Recent updates

Flide 0.1.12, released on November 10, 2026, added GLiDE, Nimble-v4-max and clef-flash models and precompiled kernels for RTX 50 series and RTX PRO 6000 GPUs; the project's Show HN on November 9 reached 124 points.

As of

07Pricing & FAQ
Is Flide free?

Yes - Flide is Open Source. Flide is free and open-source under the Apache-2.0 license and installs from PyPI. You run it on your own NVIDIA GPU; prebuilt kernels cover A10, A40, RTX 30, RTX 50, RTX PRO 6000, H100 and H200.

Is Flide open source?

Yes - Flide is open source. Flide is free and open-source under the Apache-2.0 license and installs from PyPI. You run it on your own NVIDIA GPU; prebuilt kernels cover A10, A40, RTX 30, RTX 50, RTX PRO 6000, H100 and H200.

Who is Flide best for?

Teams running Jev-style decision models for agent routing, triage, classification or LLM-as-a-judge workloads who need high throughput on their own GPUs.

Who is Flide not ideal for?

Developers looking for a coding assistant, general chat or text generation serving, or anyone without an NVIDIA GPU.

What are the best Flide alternatives?

The closest Flide alternatives on ai.dosa.dev are Copperhead, NVIDIA Personal AI Router (PAIR), WebLLM - all listed under Hardware & Edge AI.

Who makes Flide?

Flide is developed by OmniJev. It is listed in the Hardware & Edge AI category on ai.dosa.dev.

08More in Hardware & Edge
09Tags
  • Inference
  • Jev
  • Decision Models
  • Open Source
  • GPU
Track Flide in your AI stack

Favorite this tool to revisit it later, or Zap it to contribute to the public vote count.

Compare Hardware & Edge Sign in to save
Content on this page is AI-generated. Please verify details with the vendor's website for accuracy. Are you the maker? Get your Featured badge ← Back to directory