CTO planning and orchestration
A CTO agent breaks a brief into tasks, assigns specialist roles, and coordinates the run end to end so the user does not have to micromanage prompts.
AskCodi Desktop
RivendicaAskCodi Desktop is a local-first AI engineering app that plans coding work, routes tasks to models, and runs specialist agents in parallel git worktrees.
Overview
AskCodi Desktop is a local-first AI engineering app that turns a brief into coordinated work across specialist agents. A CTO agent plans the task breakdown, assigns roles, routes each task to an appropriate model, and reviews the resulting diffs before changes merge.
The product is aimed at developers and teams that want AI-assisted coding without giving up local control of their code. The site says the desktop app runs on the user’s machine, supports bring-your-own Claude or OpenAI subscriptions, and uses isolated git worktrees so agents can work in parallel with separate branches and clean merges.
Features
Model routing by task
The app routes work to different models based on the task, with examples on the site such as Haiku for refactors, Sonnet for features, and Opus only when needed.
Parallel git worktrees
Specialist agents work in isolated git worktrees on the local machine, which lets backend, frontend, test, and security work happen in parallel without stepping on each other.
Agent review before merge
Agents can review each other’s diffs and run tests before changes merge, so the workflow includes a review step rather than only code generation.
Custom agents in YAML
AskCodi supports custom agents defined in YAML, with prompts, models, tools, and guardrails bundled into one reusable definition that can be changed without rewriting code.
API and workflow integrations
The platform supports an OpenAI-compatible API and advertises integrations with IDE, CLI, desktop app, and other development workflows.
Use Cases
Turn one brief into a coordinated plan
Use AskCodi when you want a short product or engineering brief turned into a multi-step implementation plan that is split across roles such as backend, frontend, tests, and security.
Ship features with parallel agents
Use it for feature work that benefits from parallel execution, such as having separate agents handle implementation, tests, and security checks in isolated worktrees.
Control model spend on routine code changes
Use it for refactors where task routing can send simpler work to a cheaper model and reserve more capable models for harder steps.
Standardize recurring team workflows
Use custom agents when you need a repeatable review, release, or internal workflow that can be versioned in git and called from the API or supported IDEs.
Work from chat, IDEs, or CLI tools
Use the browser chat or IDE integrations when you want to ask questions, prototype with a model, or connect AskCodi to an existing editor-based workflow instead of running the full desktop swarm.
Pros and Cons
Pros
- Coordinates planning, implementation, and review in one local workflow.
- Routes tasks to different models instead of using one model for every step.
- Keeps code on the user’s machine according to the site’s local-first positioning.
- Supports parallel work through isolated git worktrees, which reduces branch conflicts.
- Offers custom agents, an OpenAI-compatible API, and integrations for broader development workflows.
Cons
- The website’s most detailed workflow is centered on the desktop app, while the browser chat is presented as better for quick questions than for heavy agent work.
- Some integration and platform details are presented at a high level on the site, so readers may need the docs to confirm exact editor, tool, and OS support for their setup.
- Pricing uses tokens for premium models, so usage costs depend on model selection and consumption rather than a single flat included plan.
FAQ
How does AskCodi Desktop work?
AskCodi Desktop uses a CTO agent to break a brief into tasks, assign specialist agents, and route work to models based on the task. The site says the desktop app and web chat are free to download, with plans that add monthly tokens for premium models across the app, chat, and API.
Is AskCodi free to use?
The pricing page says the app is free to download and use, and that plans add monthly tokens for premium models. The free tier includes free models and a one-time 100K token allowance for premium models.
Which platforms does AskCodi support?
The site lists support for Mac, Windows, and Linux on the main pricing and homepage materials, while the get-app page notes Apple Silicon is available today and Windows and Linux are coming soon there. The browser chat is also available if you do not want to install the desktop app.
What tools can AskCodi integrate with?
The integrations page shows AskCodi can be used with IDE and editor workflows such as VS Code, Cursor, Windsurf, JetBrains, and terminal or CLI setups, along with other connected tools listed on the page. The custom agents page also says agents can be called from chat, the IDE, the API, or the desktop swarm.
Can I create custom agents in AskCodi?
The site says custom agents can be defined in a single YAML file with a prompt, model, tools, and guardrails. They can be versioned in git, model-swapped, and called from the desktop swarm, the API, chat, and supported IDE workflows.
Quick Facts
- Category
- Developer Tool
- Product type
- Local-first AI engineering desktop app
- Platforms
- Mac, Windows, Linux
- Primary workflow
- CTO-led planning, parallel agents, and review in git worktrees
- Source domain
- askcodi.com
- Pricing
- Free desktop app with paid token plans for premium models