The Code Book your IDE never had.
Code Tracker AI captures every change, developer note, and architectural decision into a permanent, queryable project memory. The “why” survives the next handoff, the next vacation, the next sprint.
657 developers on the waitlist
Developers from 240+ companies · 20+ countries
+ 232 more teams across tech, education, government, tribal nations, pharma, aerospace, finance & legal
Public waitlist open since September 2025. Built by solo founder Mona Hidalgo of Mix It Hub, LLC.
Switched to token-bucket rate limiting after the Black Friday outage — the old fixed windows let traffic bursts slip through. This keeps the checkout API stable under 10x load.
Built for the JetBrains IDEs you already use — and expanding
Built for every JetBrains IDE at launch — VS Code & Visual Studio on the roadmap.
The problem
The “why” is the part that disappears first.
Handoffs
The new hire gets the repo, not the rationale. Six weeks later, they re-litigate the same rejected option.
Vacations
A long leave clears working memory. You come back to PR-842 and cannot remember what tradeoff you almost took.
Audits
Six months on, a team lead asks why a decision was made. The Slack thread is gone, the JIRA is closed, the answer is folklore.
Beyond the IDE · On the roadmap
Your project, readable on the web too.
Everything Code Tracker AI captures in your IDE will also be available in a companion web dashboard — so you can read, review, and share project knowledge without opening the editor. Great for standups, retros, handoffs, and reading Code Books anywhere.
Dashboards for your project's activity, trends, and productivity — viewable in the browser.
The plain-English story of what changed and why — ready to read or drop into a standup.
Your timestamped development log, browsable and searchable outside the IDE.
Revisit how a session unfolded — step back through the work on the web.
Read and share your docs, specs, and notes in the browser — no IDE required.
Web dashboards are planned after the JetBrains Marketplace launch. The IDE plugin comes first.
What it does
A living memory for the codebase — written by the people who actually know it.
Capture every change with its why.
As you write, Code Tracker AI logs what changed and what you meant by it — the rationale, the rejected alternatives, the constraints. Not a TODO that goes stale. A durable, time-stamped entry that ages into institutional knowledge.
Locked in a local vault.
Stored on your machine, not ours. No silent sync, no opaque cloud copy of your project memory. Recall runs locally. (At-rest encryption is on the roadmap.)
AI-powered recall, on demand.
Ask “why is this here?” — get the original decision, the constraints, the alternatives it tried, in seconds.
Docs that write themselves.
Turn the recorded decisions into the architecture overview, the runbook, the onboarding guide — already half-written.
See it in action
The whole plugin, right inside your IDE.
No dashboards to leave for. Code Tracker AI captures the session, then writes the story — what you built, why, and what's next.
A coding partner that knows your project.
Ask questions, debug, and get explanations without leaving the IDE — the chat sees your real session data, not just your prompt.
Ask & Agent modes
Ask a quick question, or hand off a multi-step task to Agent mode — right inside the editor, no context-switching.
Knows your Build Log
The chat can reference what you actually worked on — files touched, commits, session history — so answers fit your project, not a generic prompt.
Bring your own AI, or use ours
Runs on AWS Bedrock out of the box; bring-your-own-key support is on the roadmap so teams can use their own provider.
Your work, written up automatically.
Every session captured — time, files, commits, flow — then turned into an AI narrative of what you built and why. No manual logging.
Automatic session capture
Tracks active time, files edited, and flow as you work — with idle detection so breaks don't inflate your numbers.
Git-aware narratives
Pulls your commits into the story, so the write-up reflects what actually shipped, not just keystrokes.
Session replay & reports
Replay a past session or export a report — perfect for standups, handoffs, and remembering the "why" months later.
Catch issues before they cost you.
Dozens of rules across many languages, scored by severity — with a plain-English WHY, HOW, and IMPACT for every finding.
WHY / HOW / IMPACT for every finding
Not just "this is wrong" — each insight explains why it matters, how to fix it, and the impact if you don't.
Multi-language coverage
Analyzers span Kotlin, Java, Python, JavaScript/TypeScript, C#, Go, Rust, PHP, and Ruby — plus universal security, quality, and architecture checks.
Severity-ranked, not noisy
Findings are scored and surfaced by severity, so you see what actually matters first.
Stay sharp. Avoid burnout.
A built-in focus timer that nudges healthy breaks, so you keep momentum through long sessions without running yourself into the ground.
Pomodoro focus & break cycles
Work/break intervals with gentle nudges — plus lunch, stretch, and coffee reminders so long sessions stay sustainable.
Multiple watch faces
Choose from several clock styles to match how you like to see your time and progress.
Ties into your productivity data
Sessions are saved alongside your Build Log, so focus time becomes part of your work story.
From notes and blueprints to a published book — inside the IDE.
Not a scratchpad. A full Word-style writing space for specs, architecture docs, runbooks, research notes, and learning archives — rich formatting, templates, and multi-format export (PDF, DOCX, Markdown). Your documents live right next to the code they describe, versioned with the project.
Built for research teams, not just coders
Labs and research groups write far more than code — notes, protocols, blueprints, findings, and manuscripts. Code Books gives them one place to capture all of it, right where the work happens, so the reasoning stays attached to the project.
Turn your work into a book or paper
Assemble your notes and documents into long-form output and export to PDF, DOCX, or Markdown — a real path from working notes to a publishable manuscript, report, or book.
Write next to the code it describes
No switching to Notion or Google Docs. Specs, docs, and runbooks live in the same IDE as the code — templates included to start fast.
Export & share
Multi-format export (PDF · DOCX · Markdown) so your documents travel wherever your team, reviewers, or publishers need them.
In depth
Four tools, one plugin — each doing real work.
The short version is on the screenshots above. Here's what's actually under each one.
The work, written up while you do it.
- Captures each session automatically — files, diffs, Git activity, flow.
- Turns it into a plain-language AI narrative of what changed and why.
- Session replay to retrace how a piece of work actually unfolded.
- Export a session as a standup update or a report, in one click.
Issues caught in context, not after the fact.
- 62+ rules across 10 languages, scored by severity.
- Every finding comes with a plain-English WHY, HOW, and IMPACT.
- Built on IntelliJ's own PSI engine, so it reads your code the way the IDE does.
- An automated fix where one exists — not just a red squiggle.
A real document editor, inside the IDE.
- Word-style writing for specs, architecture docs, and runbooks.
- Rich formatting, headings, tables, and templates to start from.
- Export to PDF, DOCX, or Markdown — no leaving your editor.
- Lives and versions alongside the code it describes.
Momentum without the burnout.
- A built-in focus timer on the Pomodoro rhythm.
- Six-plus watch faces, rendered natively in Java2D.
- Gentle nudges to break before a long session runs you down.
- Always a glance away, never in the way.
Under the hood
Built like infrastructure, not a plugin bolted on.
- IntelliJ Platform SDK. Reads your project through the same PSI engine the IDE uses — accurate, not a text scraper.
- AWS Bedrock, model-routed. The right model for the task — fast ones for narratives, deeper ones for analysis — with a template fallback so it never hard-fails.
- Local-first vault. Your project memory stays on your own machine. No silent sync, no opaque cloud copy. (At-rest encryption on the roadmap.)
How it works
Five steps. Zero ceremony. Runs while you code.
The plugin lives in your JetBrains IDE. Capture is in-band; recall is a keystroke.
- 01
Write
You edit code in any JetBrains IDE.
- 02
Capture
The plugin reads diffs and intent as it happens.
- 03
Lock
Notes + rationale land in your local vault.
- 04
Recall
Ask any time — AI pulls the original context.
- 05
Document
Auto-roll the memory into docs and runbooks.
Use cases
Whoever you are, the same pain looks the same.
Solo developer. Scaling team. Enterprise. When the codebase outlives any single contributor's memory, recall beats archaeology.
You are the team. A four-month side project gets picked back up — and the context comes back with it.
“I left this in March. I picked it up in July and we were working again in ten minutes.”
Onboarding stops being a senior-engineer tax. New hires ask the codebase, not the calendar.
“Handing a new joiner the repo and the memory beats handing them the Slack scrollback.”
Audits, compliance, postmortems — decisions you made two years ago come with the receipt and the rationale.
“Compliance asked why we structured the write path like this. One answer: the memory entry from the day.”
Who it's for
Every industry writes software now.
Pharmaceuticals, biomedical, healthcare, energy, manufacturing, aerospace, finance, insurance, legal — these are not software companies, and every one of them ships software anyway. The mobile app. The company website. The customer portal. The internal systems nobody outside the building ever sees. The data pipelines the whole business runs on.
That code carries the same problem a tech company's does: it outlives the people who wrote it, and the reasoning behind it disappears first. Code Tracker AI runs inside the IDE those teams already work in, and keeps the reasoning attached to the code.
What these teams are building
- Customer-facing mobile apps, in Kotlin or Java
- Company websites and customer portals
- Internal tools, dashboards, and reporting systems
- Data and analysis pipelines, usually in Python
- Lab, plant, and device systems in C or C++
- Integrations between the systems they already bought
The JetBrains IDEs their engineers use
Code Tracker AI is built on the IntelliJ Platform, so it works across the JetBrains family. Visual Studio and VS Code versions are on the roadmap.
Code Tracker AI is built by Mona Hidalgo, solo founder of Mix It Hub, LLC. Mona is a solo technical founder handling engineering, design, and product herself. She started Code Tracker AI after living the problem it solves — watching the reasoning behind code disappear with every handoff, every departure, every “wait, why did we do it this way?” It’s built by someone who writes the code, not just talks about it.
Have a question or want to talk? Reach Mona directly at mona@codetrackerai.com.
Get on the list
Be there when the first build ships.
Early access goes to waitlist members first. Reserve your spot and we'll reach out the moment Code Tracker AI is live on the JetBrains Marketplace.