The software you need. Built around your rules.
Turn a business idea into a service your customers can use. Taiga carries the work through build, release and maintenance, guided by your company’s policies.
Taiga is the AI software factory from Taiga AI Oy, Helsinki, Finland, at tai.ga.
Let customers book their own deliveries.
Build the booking service, add rescheduling and keep it working with Taiga.
Discovery
Turn the request into a specification.
Your intent and published company policies shape the work.
Customers can book without calling.
Taiga builds. Your team reviews. Your pipeline deploys.
Now let them change their plans.
You put rescheduling in the queue. Taiga plans it.
The same booking. A new choice.
The reviewed change reaches the same production app.
Know how it is working.
URL checks and browser signals show how the service behaves.
A security fix joins the work.
The patch gets its own review and independent scan.
Customers keep booking. Your team moves on.
Monitoring continues
a17c942b28d153c39e264CurrentBooking confirmedDelivery rescheduled
Change delivery timeApp URL configured · browser beacon connected
Want to see it all?Continue here, at your pace.Close story
Start with the service your customers need.
Describe the business outcome. Taiga turns the conversation and your published company policies into a specification, then derives the remaining setup documents.
Your intent and your published company policies.
A self-service booking portal, built around your company’s rules.
Your team published the specification, then completed and reviewed the remaining required setup documents.
Your service opens for business.
Taiga builds against your rules. Your team reviews the result before it goes into production.
Your team reviews. Your configured pipeline deploys.
- Specification
- Queue and build
- Required review
- Configured pipeline
Your pipeline reported this successful production deployment. Taiga keeps the version and its deployment record together.
Customers want to change their plans.
Make rescheduling the next initiative. Put it in the queue: Taiga plans the change using your project context and company policies.
You choose the next improvement. Taiga plans it.
- INIT–02Let customers reschedule a delivery.Building
Keep the booking. Give customers a choice of another available time.
No work here
No work here
No work here
Build on its own · enabledMerge · required review applies
The initiative started in Backlog. Your team added it to Queue; Taiga planned the change before building it.
The improvement reaches your customers.
Your team reviews the change and your pipeline deploys it. Taiga records the version reported by your provider. The delivery portal now offers rescheduling.
Required reviews stay in force for every release.
- Reviewed
- Development
- Production
Your pipeline reported this successful production deployment. Taiga keeps the version and its deployment record together.
See how the service is working.
Availability, experience and browser-observed API errors and response times show what needs attention.
Availability, experience and API signals in one project.
Illustrative records and measurements. No live systems are connected.
The project has an App URL configured and a browser beacon installed. These signals describe the running service; repository findings are tracked separately in Maintaining.
A dependency needs a security update.Customers keep booking. Your team moves on.
Maintaining finds an urgent dependency risk. An eligible fix joins the queue, follows the configured build and review requirements, and receives an independent scan when its pull request is ready.
Repository finding → queued patch → review and independent scan.
When the fix pull request became ready, Taiga automatically started an independent scan of its branch. Here the scan no longer detects the advisory.
- Patch built
- Independent scan: advisory absent
- Required review completed
- Pipeline reported production deployment
The scan can report Fix verified, Still vulnerable or Could not verify. Required review and pipeline deployment remain separate steps.
The service is running, the improvement is live and the vulnerability has been resolved.
Illustrative workflow and metrics; no live system is connected. Eligible urgent dependency patches join the queue. Build and merge follow your configured permissions and review requirements. An independent scan checks the fix branch when its pull request is ready.
Fictional delivery business. Real Taiga interface patterns. The longer walkthroughs explain the technical decisions.
What would you put into production?
Bring an idea, an existing system or the change your team keeps putting off. We’ll work through the scope, your rules and a practical first step.
Your company’s rules are part of the build.
Your intent and published policies shape the specification. Approved technologies, data requirements and design standards guide each initiative.
Start where your software is.
You do not need a blank repository, or a finished specification, to start a useful conversation.
Evolve your software
Import your codebase. Turn its findings into prioritized work and take the next business improvement into production.
New software
Shape the user journeys and interface, then carry the idea through a controlled build into a running application.
Make AI-built apps enterprise grade
Keep the useful business workflow you built with AI. Follow it through company requirements, repairs, verification, your production pipeline and ongoing maintenance.
Keep production healthy
Turn a production finding into a repair. Follow it through delivery, deployment and independent verification.
Know the cost. Choose who operates it.
Calculate your monthly service price, choose the operating model and use the Trust Center for security and procurement questions.
Discover & Analyze from €2,000/month. Build and operate from €6,000/month including Discover & Analyze. EUR, excluding VAT. View plans and included capacity
Questions buyers ask before the first call
What is Taiga, and who is it for?
An AI software factory for regulated businesses and the public sector, where you have to show how software was built. You describe the change and publish your company’s rules. Taiga writes the specification, builds and releases the software, then keeps it maintained in production. It has carried production customer traffic since May 2026, and its paying customers are in healthcare and fintech.
Can Taiga build new software and improve our existing systems?
Both. For a new service, you work out the user flows and interface in conversation and publish them before anything is built. For an existing system, Taiga imports the code from your GitHub repository, documents what it finds and turns the findings into a backlog you prioritize. Either way, the approved change goes out through the deployment pipeline.
How does Taiga compare with our current team, consultancy or AI tools?
A coding agent is one machine inside the factory. The enterprise needs the factory around the machines. Claude Code or Codex can write the code. Taiga drafts the specification from your policies, runs the checks and keeps the requirement, the failed check, the correction and the release in one record you can inspect. With an internal team or a consultancy, people do that work between tools.
Compare building, buying and hiringCoding assistant or software factory
What does our team still need to do?
Define the outcome, publish the policies Taiga applies, connect the systems it may reach and decide which steps need your approval. Reviewing the result before production stays with your team. The specification, checks and audit trail make the work reviewable; they do not replace a competent reviewer. Your identity directory, repositories and cloud account remain yours to secure.
Who deploys, hosts and maintains the software?
Under Taiga Complete, Taiga provides the repository and runtime and is responsible for that environment’s security. In your own cloud account, Taiga builds and deploys as agreed, and once the software runs there, the runtime and the account are yours to secure. You choose the mode per connection. Whichever you choose, Taiga watches the availability, browser and API signals you configure and prepares repairs. Applying a fix is your decision.
How much does Taiga cost, and what is included?
Discover & Analyze starts at €2,000 a month for up to 10 products or repositories. With Deliver + Operate, the entry price is €6,000 a month and includes 1,000 Aurora; one Aurora is €1 of AI inference and runtime consumption. One organization, unlimited users, prices excluding VAT. When the pool runs out, consumption-driven work stops until an administrator authorizes a step-up or the next month’s pool starts. There is no per-Aurora overage invoice. Your Order Form sets the terms.
How does Taiga protect our code and data?
Customer data at rest is held in AWS Frankfurt, and no customer data leaves the EU. The models are Anthropic Claude on Amazon Bedrock in EU regions, and your content is not used to train them. Tenants are isolated from each other, and the coding agent reaches your repositories only through the GitHub App scopes you grant. We hold no ISO 27001 or SOC 2 certification yet, and no third-party penetration test has been carried out.
What happens to our code and data if we stop using Taiga?
If you connect your own GitHub, the code lives in your repository and stays there. Under Taiga Complete, Taiga hosts the repository, which does not change who owns the code. You own what the factory produces for you, under the terms of your signed agreement. A tenant owner can export the whole tenant as an archive from the product, without a support request. Under the current master terms, termination takes effect at the end of the month after the month of notice.
How can we evaluate Taiga before committing?
Bring one real project to a conversation: a new service, an existing repository or a change your customers need. We work through what Taiga would do with it and what it would cost, and references are available during the evaluation. The evidence page is public, the security answers need no inquiry, and a redacted sample pack from a real factory run is available on request.
Put your next business idea to work.
A new service, an existing system or a change your customers need. Bring it to the conversation.