Create New Project
For defining a new AI-assisted software concept.
Output package
Explore Create New Project A 14-document Project Blueprint.
FOR AI BUILDERS
Keep the project state visible while AI-assisted development moves fast
Airchor helps organise the scope, assumptions, evidence, risks, decisions, roadmaps, and handoffs around AI-assisted software work.
Airchor does not currently control or execute coding agents. It prepares structured project context for review and later implementation.
- Fragmented context
- Scope and evidence
- Roadmap and risk
- Reviewable handoff
THE AUDIENCE PROBLEM
Prompts move quickly. Project context does not.
AI-generated prototypes can accumulate files, architecture changes, hidden assumptions, and repeated prompts without a durable record of what the project is, what changed, and what should happen next.
LIVE
One workspace. Two starting points.
Use Create New Project when the software does not yet exist or still needs to be defined. Use Work on Existing Project when an application, prototype, inherited repository, or deployed product already exists and requires structured assessment or continuation planning.
Airchor helps organise the scope, assumptions, evidence, risks, decisions, roadmaps, and handoffs around AI-assisted software work.
CONTEXT → CONTROL
AI-builder orchestration motif
- Prompts→
- Decisions→
- Architecture→
- Project state→
- Governed agents
Read visual as text
- Prompts
- Decisions
- Architecture
- Project state
- Governed agents
Prompts → project state → governed AgentTasks
LIVE ROUTE MAPPING
Create New Project and Work on Existing Project
Use Create New Project when the software does not yet exist or still needs to be defined. Use Work on Existing Project when an application, prototype, inherited repository, or deployed product already exists and requires structured assessment or continuation planning.
Work on Existing Project
For continuing or assessing an existing AI-built prototype.
Output package
A 10-document Assessment & Recovery Plan.
Current boundary: Airchor organises supplied project context. It does not currently inspect or verify source code autonomously.
Explore Work on Existing ProjectCURRENT WORKFLOW
How Airchor works today
Provide structured project context.
Separate known information, assumptions, evidence, gaps, and risks.
Apply deterministic rules to readiness, confidence, evidence strength, and next actions.
Create the relevant persistent Markdown document package.
Inspect the documents, limitations, warnings, and staleness.
Download eligible documents for developer, agency, advisor, or coding-tool review.
CURRENT OUTPUTS
Structure the work before another build cycle
Airchor helps organise the scope, assumptions, evidence, risks, decisions, roadmaps, and handoffs around AI-assisted software work.
- Clarify the intended user and outcome.
- Define scope and non-goals.
- Record architecture assumptions.
- Separate evidence from inference.
- Expose missing information and risk.
- Prepare a roadmap and bounded next sprint.
- Create a developer or coding-tool handoff draft.
SYNTHETIC EXAMPLE
Prompts move quickly. Project context does not.
AI-generated prototypes can accumulate files, architecture changes, hidden assumptions, and repeated prompts without a durable record of what the project is, what changed, and what should happen next.
Project state to preserve
- The intended user and outcome
- Scope and non-goals
- Architecture assumptions
- Evidence and inference
- Missing information and risk
- The roadmap and next sprint
Current route choice
- Create New Project for defining a new AI-assisted software concept.
- Work on Existing Project for continuing or assessing an existing AI-built prototype.
Synthetic generic scenario. It is not a customer, testimonial, or product screenshot.
CURRENT LIMITATION
Current planning and assessment, not agent control
Airchor does not currently control or execute coding agents. It prepares structured project context for review and later implementation.
- No prompt or agent-session execution
- No repository inspection or code generation
- No tool, test, or build execution
- No pull-request creation or deployment
PLATFORM VISION
A control plane around coding agents
Future Airchor is designed to coordinate approved tasks, permissions, coding-agent providers, isolated execution, verification, pull requests, previews, and updated project state without replacing the underlying specialist tools.
LIVE IN MVP v1.0
Structured project state
- Planning and assessment workspace
- Create New Project
- Work on Existing Project
- Persistent projects and Saved Projects
- 14-document Project Blueprint
- 10-document Assessment & Recovery Plan
- Readiness, confidence, evidence, risk, assumptions, and next actions
- Authenticated Markdown downloads
PLANNED / VISION
Governed agentic execution
- Repository intelligence and immutable snapshots
- Grounded AI and source citations
- Approved AgentTasks and AgentPlans
- Capability and permission gates
- Local and managed execution environments
- Coding-agent adapters
- Independent verification
- Pull requests and previews
- Recurring monitoring
- Enterprise and multi-agent controls
This is the planned platform direction, not current MVP functionality.
MVP v1.0 · LIVE
Airchor helps organise the scope, assumptions, evidence, risks, decisions, roadmaps, and handoffs around AI-assisted software work.