Epho
API to run Claude Code, Codex, or OpenCode in the cloud — POST a prompt, get a streamed result back
| What is it | API to run Claude Code, Codex, or OpenCode in the cloud — POST a prompt, get a streamed result back |
|---|---|
| Pricing | Paid |
| Free tier | No |
| Platform | API |
| API | Yes |
| Best for | automating code refactoring and bug fixes, running CI/CD agent tasks |
| Domain registered | 2025 |
Data updated Sept. 19, 2026
What does Epho do?
Epho is an API that spins up AI coding agents — Claude Code, Codex, or OpenCode — in cloud sandboxes. You send a single POST request with a prompt, repo URLs, files, and MCP server configs, and Epho streams back the agent's output in real time. There's no SDK to install, no daemon to run, and no infrastructure to manage on your side. It's a straightforward way to turn a prompt into code changes without leaving your terminal or CI pipeline.
Under the hood, Epho boots a fresh environment with the agent you choose, clones your repositories, attaches any input files, and wires up your MCP servers. The whole run streams back as server-sent events — tokens, tool calls, diffs, and the final answer. You can also use the async endpoint to fire off a task and pick up the result later via webhook. Chats are durable: a sandbox pauses between turns and resumes with the same filesystem and agent session, so you can have multi-turn conversations with the same context. Epho supports bring-your-own provider keys, and some free OpenCode models don't require any key at all.
This tool is built for developers who want to automate coding tasks — think PR reviews, refactoring, bug fixes, or data reconciliation — without babysitting a local agent. It's also useful for CI/CD pipelines, cron jobs, or any script that needs to make code changes programmatically. If you can write JSON, you can use Epho. It's minimal, fast, and honest about what it does.
Key features
What makes it stand outWho is Epho for?
Who benefits most from this toolPricing
Default
- 10 starting credit
- 0 CPU rate per vcpu per second
- 0 disk rate per gib per second
- 0 memory rate per gib per second
- 0 default instance cost per second
- 0.0026 default instance cost per minute approx
- Per-second billing
- Default instance: 2 vCPU, 2 GiB memory, 10 GiB disk
- Starting credit $10 on signup
- Model tokens billed directly by provider
- Meter available via GET /credits
Trust & presence
Gallery
Click any image to enlargeAlternatives in Developer Tools
Free coding agent — use it on desktop, in the terminal, or in the cloud to build apps, edit code, and get answers without paying a cent.
Self-hosted platform to create and manage teams of AI agents that coordinate to ship code, copy, and reports
Desktop app that selects code from your project and sends it with context to web-based AI assistants like ChatGPT and Claude.
Framework and runtime for building and deploying multi-agent AI systems in your own cloud infrastructure.
AI tool that builds full-stack apps from a single conversation — describe your idea, watch the agent code it, and ship live.
Open-source framework for building voice and chat apps that work across Alexa, Google Assistant, and web platforms
Deploy and manage AI agents with a permanent URL and mid-run steerability, without managing infrastructure.
Desktop workspace for developers to run, organize, and manage AI coding agents like Claude Code and Codex