Overview
Introduction
What Hosho is, and how the three tools fit together.
What Hosho is
Hosho reviews prompts the way teams review code. It scores a prompt against research-backed checks, names what is wrong and where, and proposes a concrete fix for each finding - so a prompt change can be reviewed, argued about and merged like any other change to your system.
It is one review engine behind three surfaces, sharing one account and one API key. Which one you reach for is a question of where the work happens.
The three tools
MCP
One MCP server that adds hosho_improve_prompt (and related tools) to Cursor, Claude Code or any MCP client - the full review inline in the IDE.
Prompt Editor
A shared editor at editor.hoshoai.com where you and your team score, fix and re-run hard prompts together - your coding agent can join you there via co-presence.
Git Bot (GitHub Action)
Reviews every pull request that changes a prompt file and posts one comment with a verdict and suggested fixes. Runs in the repo's own CI - no Hosho GitHub App.
They are not tiers or alternatives. Most teams run the MCP while writing, the Git Bot on every pull request, and open the editor when a prompt needs real work.
What a review is built on
Three lines of evidence: Hosho's own research on real production prompts, published prompt-engineering research, and each model provider's documented conventions. Those are distilled into a core set of principles plus provider-specific checks, tuned by model family and by whether the target is a standard or a reasoning model.
How Hosho scores explains the four macro factors and what the numbers mean; Hosho Research is the longer argument behind them.
Start here
- Quickstart - mint a key and pick your first surface. Ten minutes.
- Already know which tool you want? Go straight to MCP, Prompt Editor, Git Bot or Co-presence.
- Point your coding agent at the agent guide for agent instructions to run any of these setups itself.