The control plane for your application stack

Connect your stack once.
Ahena configures and checks it.

Ahena connects, configures, verifies and diagnoses the services your application depends on, from authentication and payments to email, storage, DNS and AI. One control plane for development, staging, production and your AI coding agents.

Your application talks directly to your providers. Ahena manages the stack, not your traffic.

Start buildingView docs
$ ahena doctor -e production
AHENA DOCTOR  Leo · Production
✕ Production CORS allows every origin (*)
    → Run ahena configure cloudflare with your exact origins.
✕ leo.app isn't added to Resend
    → Run ahena configure resend to add it.
RESULT  12 Passed  0 Warnings  2 Failed

$ ahena apply -e production
AHENA PLAN pln_7q2m · Production (production)
~ Configure CORS on leo: https://leo.app  [CONFIRMATION_REQUIRED]
+ Add sending domain leo.app to Resend  [CONFIRMATION_REQUIRED]
? Apply this plan to Production? Yes
APPLIED pln_7q2m
  ✓ Cloudflare: Configure CORS on leo (verified)
  ✓ Resend: Add sending domain leo.app to Resend (verified)

Example session. The CLI is in beta: npm install -g @ahena/cli. Early access starts with development and staging environments; production opens later.

Before and after

Stop configuring every provider by hand, three times

Without Ahena

  1. Create provider accounts
  2. Configure authentication and callback URLs
  3. Configure CORS
  4. Add and verify email domains
  5. Create Stripe products, prices and webhooks
  6. Configure storage buckets
  7. Connect AI providers
  8. Copy environment variables by hand
  9. Repeat for staging
  10. Repeat for production
  11. Find drift by hand when something breaks

With Ahena

Connect your provider accounts once. Then:

$ ahena init --recipe saas
$ ahena add auth
$ ahena add storage
$ ahena add email
$ ahena add payments
$ ahena add ai
$ ahena plan -e production
$ ahena apply -e production
$ ahena doctor -e production

ahena add declares a capability, connects its provider if needed and plans the changes. Production waits for approval.

Connections

One control plane for the services you already use

Ahena talks to each provider's management API with credentials you connect. Your app keeps talking to them directly.

Ahena

A few of the providers

Point at a provider

Hover, focus or tap any provider to see what Ahena sets up and checks there. Ahena changes a provider only through an approved plan, then reads it back to confirm.

Every integration: what Ahena configures, checks and verifies, and how far each is verified →

Ahena Doctor

Know what's wrong before your users do

Doctor checks the live providers in an environment against what your app declares, and explains every finding. Run it from the CLI, the dashboard, or CI with ahena deploy-check to stop a deploy when production is misconfigured.

  1. DetectLive provider checks, coverage, environment separation, secrets, repository
  2. ExplainWhat, why, where, impact, and whether Ahena can fix it
  3. Fixahena doctor --fix, classified like any plan, approved by you
  4. VerifyAhena reads the provider back before calling it fixed
$ ahena doctor -e production
AHENA DOCTOR Leo · Production · health 86%
 
✓ Site URL configured
✓ Stripe live mode
✓ leo.app verified (Resend)
✓ Configured redirect URLs present
⚠ localhost present in production redirect URLs
Where: Supabase · Production · urls: http://localhost:3000/auth/callback
Ahena can fix it: Yes, after you approve a change that can't be undone
→ Remove them with ahena configure supabase --remove-redirect <url>.
 
RESULT 14 Passed 1 Warning 0 Failed

What Ahena does

Know what your app runs on, and change it safely

Stack Graph

Every capability your app uses, the provider behind it, its resources and health, per environment. The CLI, the dashboard and AI agents all read the same graph.

Doctor

Checks live providers, configuration and your repository, then explains each finding: why it matters, where, the impact and the fix.

Plans and approvals

Every change is planned before it happens. Billable, destructive and production changes need their own approval.

Verification

After applying, Ahena reads the provider back. A change is done only when the provider shows it. Unknown outcomes are reported, never guessed.

Generated integrations

Readable TypeScript in src/ahena/ that calls each provider with its official SDK. It's yours to keep, with no runtime dependency on Ahena.

Drift detection

ahena.lock records the provider state you approved. ahena drift compares it with the live providers and shows what changed outside Ahena, without overwriting it unless you approve.

MCP for AI agents

Claude Code, Codex and other MCP clients can inspect the stack, run Doctor and plan changes. Anything that isn't marked safe waits for you.

Teams and audit

Organizations with owner, admin, developer and viewer roles. Every change is recorded in an append-only, hash-chained audit log.

See Ahena build a stack

Declare, plan, approve, apply, verify

Real commands and real output formats from the Ahena CLI, on the sample project Leo.

Connect your own accounts

Link the accounts you already use (Supabase, Neon, Vercel, GitHub, Stripe, Resend, Postmark, AWS S3 and SES, OpenAI and 10 more) to each environment. Credentials are encrypted and bound to the environment they belong to, and Doctor flags test keys in production or a production secret reused anywhere else.

ahena connect stripe -e production

AI agents

Give AI infrastructure capabilities without handing it the keys

ahena mcp gives Claude Code, Codex and other MCP clients the Stack Graph, Doctor and plans, as you, under your role. It's read-only unless you start it with --allow-write.

  • Any change that isn't marked safe waits for you, in every environment.
  • You approve it in the Ahena dashboard, signed in. The agent gets a link, never the approval itself, and if nobody approves, nothing changes.
  • Provider credentials and secret values never reach the agent.
  • Every change is verified against the provider and audited as via mcp.

The MCP server runs on its own agent session, which Ahena's API refuses for approvals and secrets, so this holds even if the agent has a shell.

Ahena for AI coding agents → Giving AI agents infrastructure access safely →

  1. Agent requestClaude Code, Codex…
  2. Permissionsyour role, read-only by default
  3. InspectStack Graph, Doctor
  4. Planstored, classified
  5. You approvein the dashboard
  6. Applyone environment
  7. Verifyprovider read back
  8. Audit logvia mcp

No runtime lock-in

A control plane, not a middleman

Ahena configures and checks your providers from the side. It never sits between your application and them.

  • Readable code, official SDKs

    ahena generate writes TypeScript into src/ahena/ that calls each provider with its own SDK. It imports nothing from Ahena, and it's yours to edit.

  • Keeps running without Ahena

    If Ahena is unavailable, a correctly configured app keeps working. Planning, applying and Doctor need Ahena; serving your users doesn't.

  • Your accounts, your bills

    You connect provider accounts you own and pay them directly. Ahena never marks up usage, and disconnecting never deletes anything at the provider.

Why Ahena stays out of your runtime path → How Ahena handles security →

Who it's for

From one developer to an agency with fifty clients

  • Developers

    • Less provider setup: declare a capability, Ahena plans the provider steps
    • Generated integrations with official SDKs
    • Doctor before you ship
    • Development, staging and production kept apart
    • An MCP server for your coding agent
    Ahena for developers →
  • Engineering teams

    • Owner, admin, developer and viewer roles
    • Production changes need an admin
    • Billable and destructive changes approved separately
    • Hash-chained audit log of every change
    • Drift detection against approved state
    Ahena for engineering teams →
  • Agencies

    • One organization per client, many projects each
    • Every client's providers in one place
    • Reusable stacks from recipes
    • The same checks for every deployment
    • Read-only CI tokens per project
    Ahena for agencies →

What's ready

Available, in beta, and next

We only put on this page what Ahena does today, and say plainly what isn't finished.

  • Available
    • Dashboard: organizations, projects, environments, provider connections, Stack Graph, Doctor, plans, audit log
    • Approving and applying plans in the dashboard; changes that matter are approved only there
    • Scheduled Doctor runs and drift checks with email alerts on changes (production daily by default)
    • 20 providers, each tested against the real service within the scope its integration page states (for example test mode or read-only); plan, apply and read-back verification where a provider supports changes
    • Credentials protected by AWS KMS, two-factor confirmation for production changes, no production reveals (break-glass only), every decryption audited
    • Proving which stored secret version your app runs with (ahena verify), without the value leaving your machine
  • Beta
    • The ahena CLI (npm install -g @ahena/cli): init, add, plan, apply, doctor, drift, generate, deploy-check, verify, rotate
    • The MCP server for AI coding agents (ahena mcp)
  • Coming soon
    • Delivering secrets to your hosts (Vercel first), each target reviewed before it ships
    • Copying an environment's structure to another

See what's shipped in the changelog →

Connect your stack once. Ahena configures and checks it.

Start free with one project. Paid plans from $19 a month add projects, teammates and actions.