An AI editor for writing code—and following through when the behaviour gets complicated. Iterdock brings coding, failure investigation and API or agent rehearsal into a connected workspace, with desktop and browser experiences in development.
If you already use an AI editor, you know the value of a useful suggestion. The next question is often harder: did the change address the behaviour you actually saw?
That is the workflow Iterdock is being built around. Move from a reported problem to a repeatable check, review a proposed change, then compare what happened. Keep the code and the outcome in the same conversation.
For students, instructors, independent developers, and teams working on applications, APIs, and tool-using agents.
THE PRODUCT DIRECTION
One workspace. More of the job.
These are the capabilities we’re building toward. They describe the intended experience, not a list of features available to download today.
01 / EVERYDAY DEVELOPMENT
Make the change.
Ask questions about your codebase, plan a task, and work through inline or multi-file edits with AI assistance. Review the diff, run checks and decide what to keep.
Turn a supported failure into a scenario you can run again. Compare the behaviour before and after a change, and see which expectations passed, failed or remain unchecked.
Reproduce · Review · Compare
03 / APIS & AGENTS
Explore the difficult paths.
Rehearse timeouts, retries and unexpected API responses. For tool-using agents, examine actions and outcomes—including workflows involving more than one agent.
API behaviour · Tool calls · Agent workflows
04 / CLOUD CONTEXT
Bring context beside the code.
Investigate supported AWS, Google Cloud and Azure resources with optional, read-only assistance. See the available evidence and missing information without handing over permission to change your infrastructure.
Save useful scenarios and results so a resolved issue can become a future regression check. The direction includes sharing sanitised examples with teammates, with prerequisites and limitations intact.
Saved scenarios · Results · Team handoffs
06 / YOUR WORKSPACE
Keep your development rhythm.
Work in a full desktop editor with files, search, terminals, Git, diffs and branches. Choose from supported AI models, with an automatic selection option planned alongside manual choice.
Desktop workspace · Source control · Model choice
YOUR OWN AGENTS
Build your own agents
Create, test, run and manage your own single- or multi-agent systems in the same workspace that writes and proves code; every version passes the same Verified / Failed / Inconclusive gate before it is used.
AN ILLUSTRATIVE WORKFLOW
“Why did this request run twice?”
A booking request times out. The retry creates a duplicate. A plausible code change is a starting point; the useful next step is checking that exact behaviour.
01 / REPRODUCE
Make the problem observable.
Describe the retry and the expected result: one booking. Establish a case where the duplicate appears.
02 / REVIEW
Keep control of the change.
Use AI assistance to explore a fix. Read the diff and decide what should be applied to your project.
03 / CHECK AGAIN
Compare the behaviour.
Run the same check after the change. Keep the result and what it did—and did not—cover.
Iterdock is being built as a full desktop AI editor. Familiar editing and source-control workflows are part of the product direction. Compatibility with specific extensions will be published after testing; a familiar interface does not imply that every extension works.
Which models can I use?
We’re designing for multiple model providers and manual model choice, alongside automatic selection. The supported model list, allowances and pricing are not announced. Bring your own key is being considered for a later release.
What happens to my code and data?
Hosted AI can require selected code or context to be sent to a model provider. Before release, we’ll publish the actual data flows, provider handling, retention and available controls. There is no all-local or zero-retention promise today.
Does a passing check mean the bug is fixed?
It means the expectations in that particular check passed under the tested conditions. It does not establish that every failure is reproduced or that a change is safe in every environment. Existing tests, review and deployment checks still matter.
When can I try it, and on which platforms?
The desktop product is not available to download yet. Desktop and browser experiences are planned. Exact operating-system, processor, browser and extension support will follow release testing. No release date is announced.
DEVELOPMENT STATUS · OCTOBER 2026
A clear direction. A release still ahead.
This website is a product preview. Public installers, product API documentation and pricing are not available yet.
— The product remains in development. This website presents illustrated coding, failure-investigation and API-rehearsal workflows. These examples explain the direction; they are not recordings of a released app.
As the product qualifies for release, this page will link to real demonstrations, supported capabilities and setup documentation.