# Edge0-AI/Edge0

> An open-source streaming MoE inference framework: expert weights are offloaded from SSD on demand while a trained prerouter predicts routing ahead of time. Runs on Apple Silicon via MLX and ships with two ready-to-run model tiers (35B and 8B).

- Magnitude: 4.2 out of 10 — Steady
- Stars: 7,549 total · +101 stars measured 2026-10-29, 00:05–04:59 UTC, ≈ 322 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · License: Apache-2.0 · Created: 2026-09-08 · Last push: 2026-10-29
- GitHub: https://github.com/Edge0-AI/Edge0 · Page: https://gitnova.dev/en/r/Edge0-AI/Edge0

## Useful for

- Run a 35B MoE model on a Mac with 16 GB RAM without loading all experts into memory
- Serve an OpenAI-compatible /v1/chat/completions endpoint via edge0 serve
- Plug in a custom backend (e.g. CUDA) through the edge0/backends/base.py facade

## Why it’s here

- Star-counter measurements on 2026-10-29 (UTC), 00:05–04:59: 7448 → 7549 stars (+101). This is the change over that interval.
- Estimated end-of-day forecast: about +322 stars, using observed gains and the previous day.
- The repository is 51 days old and already has 7,549 stars.
- Top new repositories in the last 4 months: #89.
- About 84 forks a day — people are taking the code.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 1,250
- Issues and pull requests: 81
- Watchers: 596
- Average over the last week: 241 per day
- Usual pace: 168 per day
- Stars in the last hour (measured): 21
- Latest release: v0.1.0 (2026-10-21)

## Stars per day, last 30 days (oldest → newest, today is partial)

2026-09-30 … 2026-10-29: 16, 141, 217, 239, 219, 192, 218, 221, 90, 154, 206, 202, 65, 88, 214, 165, 197, 163, 156, 223, 183, 252, 247, 240, 220, 156, 301, 218, 231, 101

## Magnitude by day, last 14 days

10-12 4.4, 10-13 3.1, 10-14 4.2, 10-15 4.4, 10-16 4.4, 10-17 4.2, 10-22 3.5, 10-23 4.5, 10-24 4.5, 10-25 4.4, 10-26 4.5, 10-27 4.5, 10-28 4.2, 10-29 4.2

## Spotted in now

- Top new repositories in the last 4 months: #89

## Similar by description

1. **mnfst/awesome-free-llm-apis** — 3.2 · Steady · Language models · JavaScript · +15 stars measured 2026-10-29, 00:05–05:00 UTC, ≈ 92 by evening
   A list of LLM APIs with permanent free tiers: model providers and inference platforms with rate limits, models and links to get API keys. Aimed at developers who need free LLM access for prototypes and agents.
   Full card: https://gitnova.dev/en/r/mnfst/awesome-free-llm-apis.md
2. **mostlygeek/llama-swap** — 1.8 · Steady · Language models · Go · +2 stars measured 2026-10-29, 00:10–05:04 UTC, ≈ 13 by evening
   A Go proxy server that runs multiple local models and hot-swaps between them on demand, compatible with OpenAI and Anthropic APIs. Works with llama.cpp, vllm, stable-diffusion.cpp and other backends.
   Full card: https://gitnova.dev/en/r/mostlygeek/llama-swap.md
3. **ValerioDolci/capy-flash** — 0.8 · Steady · Language models · C++ · +0 stars measured 2026-10-29, 00:11–05:09 UTC, ≈ 2 by evening
   A C++ inference engine that runs the large Qwen3.8-Flash-Next MoE model (125B) in NVFP4 on two 16 GB consumer GPUs: routed experts are offloaded to host RAM and computed on the CPU, the rest runs tensor-parallel. It also serves the dense…
   Full card: https://gitnova.dev/en/r/ValerioDolci/capy-flash.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-10-29 05:11 UTC, updated every 30 minutes.