Parallel
Several tasks at once, without collisions
One request in development, another in testing, a third being diagnosed, each kept apart from the others.
TimoOne desktop app, a full AI software engineering team
Requirements, code and verification, handled on your computer by AI engineers who each have their own job. Every step is visible, you can stop it at any time, and every delivery comes with test evidence.
Internal use · expected to open in 2026 Q4Join the beta list
Six AI engineers, working like an engineering team
Each role does what it is good at, and hand-offs happen inside the same task, so you never have to relay context.
AI requirements analyst
01 · Requirements
Before I start: should the export include cancelled orders?
Turns a one-line request into features and acceptance criteria, and asks before guessing.
Hands you: Requirements and acceptance criteria
AI architecture analyst
02 · Architecture
This touches the order service and the reports module. Both are marked.
Reads how the existing code fits together and says what a change will touch.
Hands you: Code map and impact scope
AI software engineer
03 · Code
Built against the agreed criteria. The change sits in three files.
Writes and changes code against the agreed requirements.
Hands you: Working code changes
AI test engineer
04 · Testing
Ran 12 cases and 1 failed. Screenshots and logs are attached.
Designs test cases, prepares test data and runs the tests, rather than calling it “looks fine”.
Hands you: Test results with screenshots, logs and other evidence
AI code reviewer
05 · Review
Line 42 skips the permission check, so any user can export every order.
Reviews each change for defects and security issues, with the location and the reason.
Hands you: Review notes and an issue list
AI defect analyst
06 · Diagnosis
Found the failure: a time-zone conversion is a day off. Fix it here.
When a test fails or production breaks, finds the root cause and points to the fix.
Hands you: Root cause and suggested fix
Lines are illustrative, using an order-export request as the example.
Agentic SDLC: keep your process, make AI delivery real
Keep your repos, CI and release process; hand each step to an agent. Two quality gates hold the line, and failures return to coding with root cause and evidence.
- 1Requirements specRequirements analysisThe request is written up as a spec with features and acceptance criteria.Requirements analyst
- 2Technical design specDesign agentReads the existing code and produces the technical design spec.Architecture analyst
- 3ImplementationCoding agentImplements the change against both specs, writes unit tests, and runs the build and static scan before handing off.Software engineer
- 4Component testing and quality gateComponent test agentRuns unit tests with coverage, API tests and code review. If the gate fails, the defect analyst runs root cause analysis and it goes back to coding.Test engineerCode reviewerDefect analyst
- 5Test preparation (shift-left)ParallelTest design agent, test data agentTest cases and test data are prepared from the requirements and design.Test engineer
- 6Integration testing and quality gateIntegration test agent: API integration, end-to-endRuns integration tests with the prepared test cases and test data. If the gate fails, it goes back to coding with test evidence.Test engineerDefect analyst
- 7ReleaseRelease processOnce both quality gates pass, ships through the release process.
01Spec-driven development
Requirements and technical design are written as specs first; code and tests are baselined against them and stay traceable.
02Two quality gates
Component testing and integration testing each have a quality gate, and both must pass before release.
03Closed-loop defects
Defects that fail a gate return to the coding agent with root cause analysis and test evidence, and are regression-tested after the fix.
04Shift-left testing
Test design and test data preparation run from the specs while the code is written, off the critical path.
Illustrative flow. Timo is in internal use.
One request, start to finish
From pinning down the requirement to handing over evidence, the whole flow runs in one desktop app, with several requests moving at once.
Skills
Works the way your team works
Load CodexQA Skills, or write your coding standards and test process as skills; the AI roles follow them every time.
Background
Routine work runs in the background
Schedule nightly regressions and daily checks; they run in the background and the results wait for you.
Delivery standard
Every delivery is verified and comes with evidence
“Looks fine” is not accepted. Test results, screenshots, logs and review notes come with the delivery, and anything that failed says where and why.
Models, deployment and permissions, all your call
Your choice of model
Connect the models you choose, including models from Chinese providers.
Runs inside your network
Supports use inside a corporate network and private deployment; book an evaluation for your setup.
On Windows and macOS
The desktop app runs on Windows and macOS, so everyone can use the computer they already have.
Automated, with permissions in your hands
Running commands, reaching the network and using credentials are off by default and granted by you one at a time. Nothing runs outside what you grant, and you can stop a task at any point.
Where to start depends on what you need
- Available
CodexQA Skills
Eleven open-source verification skills for the coding agent you already use (Cursor, Claude Code, Codex, OpenClaw). Free, running locally.
Try the skills free - Internal · 2026 Q4
Timo
A desktop app with a full AI engineering team: requirements, code, tests and review, with several tasks moving at once.
This page - Internal · opening from 2026 Q4
CodexQA platform
Seven modules for teams, including test cases, API and UI automation and cloud devices. Private deployment can be evaluated.
About the platform
It is not ready for public download. Progress stays public.
The desktop is used internally for now. The dates below are the current plan, not promises; completed work, delays and changed plans remain on the record.
- 2026 Q3In Progress
Agent 桌面端内测
macOS / Windows 内测包，连接用例管理与接口平台；用例生成、接口用例两条链路先跑通。
- 2026 Q4Planned
Agent 公测
macOS / Windows 公开下载；接入 UI 自动化平台，浏览器内录制与自愈脚本。
- 2027 Q1Planned
Agent 1.0 正式版
五条能力链路稳定；技能库、五个平台全部打通。
- 2027 Q2Exploring
Agent Linux 版与无头 CLI
在 CI 容器里运行 agent，无桌面依赖。
Questions
What is Timo?
Timo is a desktop app with a full AI software engineering team. AI requirements analysts, architecture analysts, software engineers, test engineers, code reviewers and defect analysts work together in one app, taking work from requirements through code to verified tests, and every delivery comes with test results, screenshots, logs and other evidence.
Can I download it now?
Not yet. The desktop app is currently in internal use, expected to open in 2026 Q4. You can book an evaluation to see the full flow on your own repository, or join the beta list to hear when it opens.
How is it different from coding agents such as Codex or Cursor?
General coding agents centre on writing code; how to break down the requirement, how to test it and whether the result counts is usually left to you. Timo makes requirements analysis, testing, code review and defect analysis dedicated AI roles, so one request goes from breakdown to verification in the same task and ships with evidence. It also supports models from Chinese providers, corporate networks and private deployment.
Which models does it support? Can it run inside a corporate network?
You can connect the models you choose, including models from Chinese providers. It supports use inside a corporate network and private deployment; book an evaluation for your setup.
Which operating systems does it support?
Windows and macOS.
Does the AI get shell, network or credential access by default?
No. Those permissions are off by default and granted by you one at a time; nothing runs outside what you grant. Every step is visible while a task runs, and you can stop it at any time.
How does it relate to CodexQA Skills?
CodexQA Skills are eleven open-source verification skills for the coding agent you already use, free today. The desktop app can load CodexQA Skills as well as skills your team writes. Before it opens, the skills are a way to try requirement analysis, code review, defect scanning and browser runs.