We give Artificial Intelligence more space to act

Free the hours Take the work Raise the bar

We help professionals, teams and companies step into the world of Daidala AI and take the full advantage that artificial intelligence gives. Click the button below and book a free consultation - we will talk about your needs and the challenges in front of you.

// turning point

Recognize yourself?

I am a professional

  • I still run out of time
  • I know I am not using AI to its full potential
  • I still do a lot of the work by hand
  • I worry about burning out

I run a business

  • I run a company and employ people
  • I know my employees are not using AI to the full
  • Better automation is still a plan for the future
  • I want to get more free time back
Daidala AI

I use Daidala AI

I have more free time

I am more rested and happier

I deliver work of better quality

My projects are delivered faster and cheaper

I plug in one or more of my own AI subscriptions

// what it is

Daidala AI is the heart of a general-purpose IT ecosystem

Like the bloodstream in a body, it joins three pillars into one coherent whole: project management software, a code repository and artificial intelligence. It organizes an isolated flow of work and data between these components, so that AI carries out the tasks defined in the projects.

The code repository on that list does not confine Daidala to the IT industry. AI agents use it for many things in their everyday work: they generate handy scripts, code and snippets, define recurring jobs, keep static websites in it and process data deterministically. In software development there is more on top: storing code, pipelines and deployments, and connecting runners.

// who it is for

You already use AI, and you still explain everything from scratch

AI speeds up single tasks, but keeping an eye on it stays with you: checking, explaining, describing the same thing once again. What is missing is not a better model - it is a way of working in which AI has its own projects and its own tasks.

For professionals

  1. You run your own business and deliver work for clients. There is a lot of it, and you know you cannot keep this pace for long.
  2. You already pay for subscriptions with the popular AI providers, and the work goes a bit faster. But the machine needs a lot of explaining, so attention goes out where it was meant to come back.
  3. What is missing is automation - the repeatable steps are still done by hand.

Write to us. We will hear you out, without rushing you, and help you set Daidala AI up around your own work - so that more of your time comes back to you.

For businesses and teams

  1. You run a company, or a team delivering your own and your clients’ projects. The pace of the work has not moved in a long time, even though the market has picked up.
  2. Your team started working with AI, and the pace went up. There is also more to do than before: watching it, checking it, explaining it, and then writing the project up.
  3. What is missing is a way of working in which AI has clearly defined tasks inside your projects, and automation fits into the steps that already run.

Write to us and tell us where the work with AI gets stuck. We will help you put Daidala AI where it makes a visible difference to your company.

// what it costs

You pay one flat price a month for your own instance

Prices are net and the same every month. We sell to businesses: a company from Poland pays Polish VAT on top, a company from another country accounts for the tax itself. The variant for teams is still being built. Neither price includes access to AI models; the “AI access” section says what you connect yourself.

On sale now

Your own instance

249 EUR net, per month

Instance specs

  • A separate server just for you: 8 vCPU, 16 GB RAM, 320 GB NVMe storage.
  • The server stands in a data centre in the European Union.
  • Your instance at its own address: you pick the name when you buy.

In the price

  • Your own Daidala AI instance, for one company, on a server of its own.
  • Project management, a code repository and artificial intelligence, joined into one whole.
  • Paid a month in advance. The contract runs month to month, with no minimum term; you turn renewal off in your account whenever you want.
  • Trial period: after one or more individual consultations. Book a free consultation.

Not in the price

Access to AI models. You connect your own paid subscription - the “AI access” section says how.

Subscribe to Daidala

In preparation

Instance for a team

999 EUR net, per month

Not on sale yet.

Instance specs

  • A separate server for the whole team: 16 vCPU, 64 GB RAM, 360 GB NVMe storage. Dedicated CPU cores, not shared with other servers.
  • The server stands in a data centre in the European Union.
  • Your instance at its own address: you pick the name when you buy.

What the price is to cover

  • The same instance, but for the whole team: several people from one company hand out the work in it.
  • Each of them connects their own AI subscription. The work they hand out runs on their own account with the provider - as the providers’ terms require.
  • Not settled yet: for how many people, and how it is bought.

Not in the price

Access to AI models. Each person on the team connects their own paid subscription - the “AI access” section says how.

// AI access

You bring your own paid access to AI models

The agents work by using AI models, and the price of Daidala does not include them. You connect your own paid access - a subscription on your own account with an AI provider we support - and you pay that provider yourself. Until you connect it, your instance runs but the agents do not do the work you buy Daidala for.

Why we built Daidala for a flat plan

AI providers sell their models in two ways: per token sent and received, the way programs usually pay for AI through the provider’s API, or as a flat monthly plan. We built Daidala for the plan only. Your instance takes one kind of access: a flat monthly plan we support, on your own account at your AI provider.

The reason is how agents work. At every step, an agent sends the model everything it has read and done so far, so the number of tokens grows with every step, not with the result. A team of agents working steadily goes through hundreds of millions to billions of tokens a month. Billed per token, that work would cost several times to more than ten times the price of a flat plan - in peak months more than twenty times.

not how Daidala works

Billed per token

A per-token bill that grows with every day of steady agent work. Drawn to the 8-month average from the figure below: about 1,900 USD a month, spread evenly over the days, against a plan for about 100 USD on the same account. The bill passes the price of the plan on day 2, is three times it after five days, six times after ten, nineteen times at the end of the month. That is the average: the peak month of the same measurement comes out at twenty-eight times.

  • You learn the cost after the month, from the bill.
  • Every step of every agent adds to it.
  • More work means a bigger bill.
how Daidala works

A flat plan

The price of the plan stands at the same height on every day of the month: about 100 USD in the same measurement, paid up front to your AI provider. It does not change with how much work the agents do within the plan’s limit. Once the limit is used up, the agents wait for the window to reset, and the bill stays where it is.

  • You know the cost at your provider before the month starts.
  • Work within the plan’s limits adds nothing to it.
  • More work means one more plan, at a known price.

Published measurements: a month of agent work billed per token, against the plan it was done on

Figures from other people’s published logs and from the providers’ public price lists, read on 1 Oct 2026. This is not a measurement of Daidala. Each row is one month: what the same work would have cost at the provider’s per-token prices, and the price of the flat plan it was actually done on. The saving is the difference between the two - what did not go to the AI provider that month; Daidala’s own price is not part of this sum. Bars are scaled within each row.

  1. One developer, a coding agent every working day - the peak month201 sessions across more than 45 projects · ksred.com, log analysis, Feb 2026 (updated Aug 2026)
    per token5,623 USD
    plan200 USD
    saving5,423 USD
    28×the plan price
  2. The same developer, averaged over 8 monthsabout 10 billion tokens, over 90% of them read from cache · ksred.com, the same logs
    per tokenabout 1,900 USD
    planabout 100 USD
    savingabout 1,800 USD
    19×the plan price
  3. A team of agents on one account: 1 billion tokens a month, in the same proportionsOur own arithmetic, not a measurement: the provider’s public per-token prices and plan prices, 1 Oct 2026
    per token1,100-1,600 USD
    plan100-200 USD
    saving900-1,400 USD
    5.5-16×the plan price
  4. A coding agent on a 200 USD plan: the same tokens at per-token pricesDX, AI coding assistant pricing and ROI guide, Jun 2026
    per token600-1,500 USD
    plan200 USD
    saving400-1,300 USD
    3-7.5×the plan price
  5. One coding agent for part of the working day: 20-60 million tokens a monthcache reads not counted · Office Claws, Apr 2026
    per token80-400 USD
    plan20 USD
    saving60-380 USD
    4-20×the plan price
  6. A person working with the agent by hand, not a team of agentsthe provider’s own figures for its coding agent · provider documentation, read 1 Oct 2026
    per token150-250 USD
    plan100-200 USD
    savingclose to zero
    0.75-2.5×roughly even

A plan has a limit, and no provider publishes how many tokens fit in it, so the comparison is a range, not one number. The last row shows where the difference comes from: a person working by hand uses far fewer tokens than agents working all day, and only steady work tips the balance towards the plan. Per-token prices and plan limits changed several times in 2026 at both providers; the amounts above are as published on the dates given.

Pool several, not one

Add several of your own subscriptions to a single pool. Daidala sends work to whichever still has limit left. When every one of them has used up its window, the agents wait for the reset and then carry on - the work waits, it is not lost.

A flat plan at your provider

Your instance takes one kind of access only: a flat monthly plan we support, at your AI provider. That plan is your cost at your provider, on top of the price of Daidala. Its limit is shared with your own use of the same account, and how much work fits in one plan is set by your provider.

oauth pool — your subscriptions
● flat monthly plan at your provider · on your own account
● window used up on every account · agents wait for the reset, then carry on