jmerelnyc/Talos
Let your GPU earn its keep serving LLM jobs
This client sells your spare GPU cycles to the Talos network, running open-source LLM inference locally via Ollama and paying you for uptime and served jobs.
Feature · 05 Jul 2026
Renting Out Your GPU, One Ollama Job at a Time
Talos worker turns idle desktop GPUs into inference nodes, betting that LLM serving will look more like an Airbnb for gaming rigs than a hyperscale data center.
Read the in-depth article →
Collecting fresh signals — velocity needs a few days of history.
collecting data…
star history
What it does This Python client turns your local machine into a node on the Talos network. After pairing with a device code, it opens a WebSocket to the Talos gateway and fields open-model inference requests, running them through your local Ollama installation. It tracks uptime and completed jobs to calculate your share of usage revenue, and hosts a local dashboard where you can dial back how much compute you donate.
The interesting bit The repo serves two masters. Alongside the worker daemon, it distributes an SDK and ready-made editor integrations for Cursor, VS Code, Claude Code, JetBrains, Zed, and Aider—so it is both the supply-side client and the demand-side onboarding kit.
Key highlights
- Delegates all inference to a local Ollama instance; the worker just brokers jobs and returns results
- Revenue is tied to actual served jobs and connection uptime, not raw availability
- Local web dashboard includes an allocation slider to limit concurrency and duty cycle
- Auto-detects NVIDIA GPUs, with CPU as a fallback option
- Bundles setup scripts and config snippets for half a dozen editors and IDEs
Verdict Consider it if you have idle GPU capacity and want to monetize it through an open-model marketplace. If you are simply looking for a standalone local LLM server or want to avoid external accounts and revenue-sharing platforms, this is not the right tool.
Frequently asked
- What is jmerelnyc/Talos?
- This client sells your spare GPU cycles to the Talos network, running open-source LLM inference locally via Ollama and paying you for uptime and served jobs.
- Is Talos open source?
- Yes — jmerelnyc/Talos is open source, released under the MIT license.
- What language is Talos written in?
- jmerelnyc/Talos is primarily written in Python.
- How popular is Talos?
- jmerelnyc/Talos has 587 stars on GitHub.
- Where can I find Talos?
- jmerelnyc/Talos is on GitHub at https://github.com/jmerelnyc/Talos.