by carloslfu · Codex Skill · ★ 483
slotstream Run a 105 GB AI model on a Mac that can't hold it. Slotstream runs Qwen3.8-Flash-Next, a 125-billion-parameter open model, on Macs with 16 to 64 GB of memory. It keeps most of the model on the SSD and loads the parts it needs as it writes. Our 48 GB M5 Pro measured 15.86 tokens per second at a 22 GB memory target (how it was measured). Chat with it, ask it about pictures, or code with it: starts Claude Code on the local model, and Codex, Pi, opencode and Hermes work the same way.
| Stars | 483 |
| Forks | 40 |
| Language | Swift |
| Category | Codex Skill |
| License | MIT |
| Quality Score | 51.92/100 |
| Open Issues | 74 |
| Last Updated | 2026-10-25 |
| Created | 2026-08-28 |
| Platforms | claude-code, cli, codex |
| Est. Tokens | ~34661k |
These tools work well together with slotstream for enhanced workflows:
Explore other popular codex skill tools:
slotstream is Run a 105 GB AI model on a Mac that can't hold it. Slotstream streams Qwen3.8-Flash-Next (125B mixture of experts) from your SSD and caches the busiest experts in memory, so it runs on Macs with 16 to. It is categorized as a Codex Skill with 483 GitHub stars.
slotstream is primarily written in Swift. It covers topics such as apple-silicon, claude-code, inference-engine.
You can find installation instructions and usage details in the slotstream GitHub repository at github.com/carloslfu/slotstream. The project has 483 stars and 40 forks, indicating an active community.
slotstream is released under the MIT license, making it free to use and modify according to the license terms.