Point an existing OpenAI-compatible chatbot at Byto to run it on Google's Gemini models instead of OpenAI.
Deploy Byto on Cloud Run to give a chat app a production ready connection to Vertex AI.
Use Byto's live traffic logs to debug token usage, errors, and provider responses without touching app code.
Run Byto locally with Docker to experiment with Vertex AI's free tier for a personal chat project.
| trfhgx/byto | 1296018244/grok-manager | 787a68/hubproxy | |
|---|---|---|---|
| Stars | 16 | 16 | 16 |
| Language | Go | Go | Go |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 4/5 | 3/5 | 3/5 |
| Audience | developer | ops devops | ops devops |
Figures from each repo's GitHub metadata at analysis time.
Requires a Google Cloud project with Vertex AI access and Go 1.22 or newer, or Docker to run the packaged service.
Byto is a small Go service that sits between chat applications and Google's Gemini models. Many chat tools and apps only know how to talk to OpenAI's API format, and Byto translates those requests into calls that Google's Vertex AI can understand, so you can point an existing OpenAI-compatible app at Byto instead of rewriting it to speak Vertex AI directly. People use it for roleplay chatbots and other chat experiences, and developers also use it to test and run AI applications against live production traffic, including experiments on Vertex AI's free tier. Byto handles the Google Cloud side of things in one place: managing access, tracking which models are available, streaming responses, caching, and reasoning controls. It can run as a single small service on your own machine, inside Docker, or on Google Cloud Run. Once running, Byto exposes familiar OpenAI-style endpoints such as chat completions, model listing, and a health check, plus some of its own, including endpoints for async chat jobs and Vertex specific caching. It keeps structured logs of requests, provider responses, token usage, and errors so you can see what is happening with live traffic without changing your application's prompts. There is no default model: if a request does not specify one, Byto returns an error rather than guessing. Which models are allowed comes from a bundled configuration file, aliases, or an environment setting that permits any Gemini model, and Byto checks each candidate model against Vertex on startup to confirm it actually works before enabling it. To get started, you can download a prebuilt binary from GitHub Releases for macOS, Linux, or Windows, or clone the repository and build it from source using Go, with setup scripts provided for Mac and Linux (Bash) and Windows (PowerShell). A Google Cloud project with Vertex AI access is required. Docker Compose is also supported for a one command startup. The project includes automated tests, including live checks against real Google Cloud accounts, and load tests built with k6. Byto is released under the MIT License.
Byto is a Go gateway that lets OpenAI-compatible chat apps talk to Google's Vertex AI Gemini models without rewriting the app.
Mainly Go. The stack also includes Go, Vertex AI, Gemini.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 1h+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.