E

Epho

Agents as an API — spin up Claude Code, Codex, or OpenCode in the cloud with a single POST.

Usage-based per-second billingBYO model provider keysSupports Claude Code, Codex, OpenCode$10 starting creditSSE streaming API
Category
AI developer tools / agent orchestration API / cloud sandbox abstraction
Business model
Usage-based infrastructure metering with BYO model keys; no token markup visible.
Target audience
Developers, DevOps/CI automation teams, and technical founders who want coding agents triggered by scripts, cron, GitHub Actions, or apps without managing sandboxes.
Opportunity score
58/100

A clean, well-positioned developer API with a real use case, but it sits in a crowded, low-margin, easily bundled infrastructure niche.

Founder verdict
MAYBE

Epho is a beautifully simple API for coding agents, but it is an infrastructure race, not an obvious standalone business yet.

What is Epho?

Epho is a developer-facing API service that runs coding agents (Anthropic's Claude Code, OpenAI's Codex, and open-source OpenCode) in fresh cloud sandboxes. A caller sends one JSON POST to /api/v1/chat with a prompt, optional repos, input files, MCP servers, env vars, and provider key; Epho boots the sandbox, clones repos, starts the chosen agent, and streams tokens/tool calls/diffs back over SSE. It also supports async jobs, durable multi-turn chats with session snapshots, persisted event replay, and presigned artifact downloads. Observed pricing is metered per-second for compute, memory, and disk only; model tokens are not billed by Epho because customers bring their own provider keys. No revenue, growth, team size, or customer proof is disclosed on the site.

Get the full Epho playbook

The complete reverse-engineering — what to copy, what to avoid, and exactly what to build instead.

Unlock full startup intelligence — $39.99