Prompts deserve version control.
mowa is the version history, eval suite, and audit log for the prompts running in your product. Every change — from a GitHub push or an in-app edit — auto-diffed against the version before, with regressions flagged before they ship.
Free forever for up to 3 prompts · No credit card required
Why mowa
Treat prompts like the production code they are.
Track every change.
Every prompt is versioned. See diffs against any prior version. Trace exactly when a behavior changed, and why.
- Full edit history with diffs
- One-click rollback
- Branch and review like code
onboarding_welcome
- v132 min ago · PM
- v1214 days ago · engineer
- v111 mo ago · PM
- v102 mo ago · engineer
Catch what changed. Catch what broke.
Every new version — from GitHub or the editor — is auto-diffed against the one before. Replay your eval suite across both and block the promote when the numbers move the wrong way.
- Auto-compare every GitHub sync
- Replay evals across both versions
- Block promotion on regressions
eval · v12 vs v13
v12
82.1%
v13
86.3%
+4.2 pp · 0 regressions across 240 cases
Know who shipped what.
Every change is attributed. Author, timestamp, reason, and the diff itself, captured in a tamper-evident log.
- Author attribution on every edit
- Compliance-ready audit log
- Exportable for SOC2 / ISO27001
audit log
- 14:22PM promoted v13
- 14:18eval suite passed (240 cases)
- 14:14PM proposed v13
The product
One workspace. Every prompt. Every version.
See your full prompt library, the diff between any two versions, and the eval results, without leaving the page.
onboarding_welcome
src/prompts/onboarding.ts
error_fallback
src/prompts/errors.ts
trial_upgrade_cta
src/prompts/cta.ts
support_triage
src/prompts/support.ts
email_nudge_day3
src/prompts/email.ts
Proposed change
v12 → v13
“Replace simple and fast with personal and unhurried.”
v12 accuracy
82.1%
v13 accuracy
86.3% ↑
PM · proposed v13 · 2 min ago
Compare & catch
Every change, compared. Every regression, caught.
Push a prompt edit to GitHub, save in the editor, or accept a PM's proposal — mowa snapshots a new version, diffs it against the one before, and replays your eval suite across both. A 4-point accuracy drop catches itself, not a customer.
- Side-by-side diff on every change — GitHub or in-app
- Replay evals across both versions automatically
- Block promotion when the numbers move the wrong way
- One-click rollback to any prior version
support_triage · v18 → v19
GitHub sync · 2 min agoYou are a friendly
- support agent who responds in 2 sentences.
+ support agent. Be concise. Never apologize.
Escalate billing issues.
eval delta · 240 cases
regressionPromotion blocked. v19 regressed tone by 3.1 points.
How it works
From Slack thread to shipped prompt, in one place.
Set up in five minutes. Use it the same week. No new tools to wedge into the team's workflow.
Connect GitHub
Point mowa at your repos. We scan and import every prompt automatically. No tagging, no config, under 60 seconds.
Invite your team
Owners, editors, viewers, and PM roles. Non-technical teammates can read, suggest, and review without a GitHub seat.
Propose, test, ship
Edits become versioned proposals. Run evals against prior versions. Promote when the numbers move the right way.
Audit, always on
Every change carries an author, timestamp, and reason. The audit log is queryable, exportable, and tamper-evident.
MCP server
Edit prompts from wherever you work.
Every workspace ships with an MCP server. Connect it to Cursor, Claude Desktop, or any MCP-compatible client. Your AI agent can read prompts, list versions, and propose changes that flow through the same review process as your team.
$ _
Pricing
Simple pricing. No surprises.
Start free, upgrade when your team grows. Annual billing available.
Free
Forever. No credit card.
- Up to 3 prompts tracked
- Version history
- Up to 3 team members
- Read-only MCP server
Pro
Most popularUp to 5 users · unlimited prompts
- Unlimited prompts
- Side-by-side eval runs
- Full audit log with author history
- Team roles · PM, editor, viewer
- GitHub PR integration
- Read + write MCP server
- Priority support
Need SSO, SCIM, or a custom audit log retention? .
Ship prompts with the rigor they deserve.
Connect a repo and have your prompt library versioned, tested, and audited in under five minutes.