Airchor MVP v1.0
The first stable public authenticated Airchor release.
User value: A live product that can be opened, used, left, and continued through a persistent project workspace.
Explore MVP v1.0AIRCHOR ROADMAP
Airchor MVP v1.0 is live today as an invitation-based planning and assessment workspace. The roadmap extends that foundation into repository-aware project intelligence, approved agent tasks, controlled execution, independent verification, and reviewable delivery.
This roadmap communicates direction and sequence. Planned work may change, and no delivery date is promised unless published separately.
Sequence communicates dependency, not dates or percentage complete.
Airchor MVP v1.0 provides a real authenticated workspace for structuring new and existing software projects, preserving project state, generating document packages, and preparing clearer next actions and handoffs.
The first stable public authenticated Airchor release.
User value: A live product that can be opened, used, left, and continued through a persistent project workspace.
Explore MVP v1.0Invited users can sign in to a personal Airchor workspace.
User value: Protected project access and persistent user sessions.
Boundary: Public self-sign-up is not currently available.
Sign inEach authorised user receives a workspace boundary for persistent Airchor projects.
User value: Projects and generated documents remain associated with the authorised workspace.
Boundary: Team invitations and complete organisation collaboration are not part of MVP v1.0.
Successful project creation, submitted workflow stages, evidence, findings, and generated documents are saved.
User value: Users can leave Airchor and return to continue their work.
Boundary: Unsubmitted form text is not automatically saved.
Returning users can use Saved Projects, Open Latest Project, and Continue Project.
User value: The project remains a durable workspace rather than an isolated chat transcript.
Boundary: Continue Project opens the project overview rather than the exact last page visited.
A complete route for defining a new software concept, scope, features, architecture assumptions, roadmap, and first handoff.
User value: A structured path from early idea to a reviewable Project Blueprint.
Current proof: 14 deterministic Blueprint documents.
Explore Create New ProjectA structured route for organising reported repository context, evidence, findings, recovery planning, and handoffs.
User value: A clearer assessment and next-step framework for an existing application or prototype.
Current proof: 10 deterministic assessment and recovery documents.
Boundary: Airchor does not currently inspect or verify repository code automatically.
Explore Work on Existing ProjectAirchor displays readiness, confidence, evidence strength, risk, assumptions, missing information, and recommended next actions.
User value: Users can see where project context remains incomplete, uncertain, conflicting, or risky.
Boundary: These indicators are advisory and do not certify technical, security, or Production readiness.
Review readiness informationSaved project inputs are transformed into consistent Markdown documents through rules and templates.
User value: The same saved input and application version produce a stable reviewable result.
Boundary: The current generation engine is not a Production repository-analysis AI system.
Eligible generated documents can be downloaded individually as exact persisted Markdown.
User value: Users can review and use the outputs with developers, agencies, advisers, or coding tools.
Boundary: PDF, DOCX, ZIP, public-share links, and combined hosted packages are not currently supported.
Review download supportThe current development focus is to align the public website, trust structure, product authority, and future technical foundation with the released Airchor product.
Replacing the legacy website with a live-product gateway, current-product explanation, trust structure, and agentic-platform narrative.
User value: A clear path from airchor.com to the live application, with accurate current capability and roadmap information.
Boundary: The website transformation does not itself add new application features.
Removing stale launch-era language and aligning current product, access, roadmap, and limitation claims.
User value: Visitors receive an accurate explanation of what Airchor does today.
Creating a service-wide trust, privacy, terms, contact, and operator structure for the live product.
User value: A clearer public explanation of product boundaries, responsible use, and approved legal information.
Boundary: Substantive legal, privacy, retention, deletion, and operator statements require formal approval.
Separating current product authority from historical MVP-development documents and preparing a smaller active source set.
User value: More consistent product decisions and fewer contradictory public claims.
Replacing interface-led product proof with bespoke diagrams for orchestration, context, approval, capabilities, execution, verification, evidence, and delivery.
User value: The website’s visual identity reflects the platform direction while keeping the live MVP foundation accurate and secondary.
Preparing the product models and architecture required for exact repository identity, source facts, evidence, and snapshots.
User value: A future path from reported project context toward repository-grounded assessment.
Boundary: Repository analysis is not currently available in MVP v1.0.
Defining safe, bounded, traceable handling for project documents and future source evidence.
User value: A foundation for richer project context without obscuring provenance or data boundaries.
Boundary: Broad multi-document, PDF, DOCX, image, archive, and source-code analysis is not currently available.
The planned platform extends Airchor from structured project context into repository-aware intelligence, approved agent tasks, controlled execution, independent verification, and reviewable delivery.
Connect an approved repository and preserve the selected provider, repository, branch, and exact source state.
User value: Project conclusions can be tied to the real repository rather than only reported context.
Boundary: Private repository connection and content analysis are not current MVP features.
Preserve the exact commit, tree, manifests, selected files, and retrieval state used by a project assessment or task.
User value: Plans and findings remain linked to a specific source state even when the repository changes later.
Turn repository facts, project documents, evidence, and findings into reviewable source-linked project context.
User value: Users can distinguish observed facts from assumptions, user reports, and generated interpretation.
Use approved source evidence to support structured synthesis, planning, and recommendations.
User value: AI-assisted outputs can cite the project evidence they rely on and expose uncertainty.
Boundary: Grounded Production AI is not a primary live MVP capability.
Convert a finding, roadmap item, or handoff into a bounded unit of work with objective, scope, exclusions, acceptance criteria, and dependencies.
User value: Execution begins from reviewed project context rather than an empty prompt.
Freeze the approved task plan, expected files, commands, checks, risks, and stop conditions before execution begins.
User value: Scope changes become visible and require a new review rather than silently modifying the approved plan.
Define exactly what an agent may do, where, for how long, and with whose approval.
User value: Repository, file, terminal, network, secret, pull-request, deployment, and database actions remain bounded and inspectable.
Boundary: The planned model is default deny and task scoped.
Show the approved objective, current phase, agent permissions, events, changed files, tests, verification, artefacts, cost, and recommended next action.
User value: Users can understand what is happening without relying only on provider-specific logs.
Boundary: This will be a project-attached experience, not a separate generic developer console.
Run approved tasks inside a registered local environment with short-lived leases, bounded capabilities, event reporting, cancellation, and cleanup.
User value: Source can remain within the user's environment during the first execution phase.
Boundary: Production credentials and unrestricted system access are excluded.
Connect the Airchor task, permission, event, and verification model to a first coding-agent provider through a provider-neutral interface.
User value: The underlying agent can change without rewriting the Airchor product model.
Create an exact-base branch or worktree for bounded changes without modifying the user's active branch.
User value: Implementation can be inspected, cancelled, retried, or discarded without contaminating the primary checkout.
Run allowlisted inspection, test, typecheck, lint, and build commands within the approved environment and budget.
User value: Task progress is supported by observed command results rather than model claims.
Boundary: Dependency installation, network access, destructive actions, and Production operations require separate controls.
Check the final diff, approved scope, tests, build, artefacts, and delivery state outside the coding-agent session.
User value: A coding agent cannot mark its own work complete without separate evidence.
Preserve task plans, approvals, events, command results, diffs, test reports, verification results, and completion records.
User value: The project keeps a durable history of what was approved, attempted, observed, and delivered.
Deliver a verified result through a reviewable branch and draft pull request after separate approval.
User value: External delivery remains visible and reversible.
Boundary: Merge and protected-branch actions remain outside the initial workflow.
Create or recognise a preview linked to the exact verified commit and persist its status in Airchor.
User value: Reviewers can inspect a bounded result without promoting it to Production.
Boundary: Production promotion is excluded from the initial roadmap.
Extend a reviewed finding into an approved task, isolated change, independent verification, and draft pull request or preview.
User value: Move from assessment into a reviewable repair without losing the evidence and reasoning behind it.
Extend an approved Project Blueprint into bounded repository creation, scaffolding, verification, and reviewable delivery.
User value: Carry the original project scope, assumptions, and acceptance criteria into implementation.
Boundary: This follows the first complete WOEP repair workflow.
The long-term Airchor vision is a provider-neutral control plane that coordinates project intelligence, permissions, agents, execution, verification, delivery, and updated project state across different tools and environments.
Compare approved project snapshots over time and surface meaningful changes, new risks, stale assumptions, and recommended next actions.
Boundary: Monitoring should begin as consent-based read-only comparison, not automatic repair.
Support more than one coding-agent provider through one task, permission, event, artefact, and verification model.
User value: Users can choose tools without losing Airchor's governance and project history.
Extend repository and delivery workflows beyond the first source-control provider.
User value: Airchor can support different development environments without changing its core control model.
Add controlled cloud sandboxes alongside local and enterprise-managed runners.
User value: Users can select the execution boundary that matches their source, security, and operational requirements.
Execute approved tasks inside customer-controlled infrastructure with organisation policy and runner trust controls.
Add organisation access, role separation, approval chains, policy administration, retention, and audit export.
Boundary: Team and enterprise collaboration are not current MVP capabilities.
Require different qualified users or roles for plan approval, privileged action approval, verification, and delivery.
Coordinate specialist or dependent agents while preserving scope, ownership, task dependencies, permissions, conflict handling, and evidence.
Boundary: Multi-agent work is a long-range capability and should not be implied by the first agent integration.
Show project health, active tasks, approvals, verification, risk, and delivery across multiple projects or teams.
Move from project context through planning, execution, verification, delivery, and updated project state while preserving the reasoning and evidence behind each step.
The current product is a planning and assessment workspace. The following capabilities are not live today.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
This capability is not part of Airchor MVP v1.0.
ROADMAP QUALIFICATION
This roadmap communicates direction and sequence. Planned work may change, and no delivery date is promised unless published separately.
These limitations should remain visible until the corresponding capability is released and independently verified as part of a later product version.
PRODUCT UPDATES
Receive occasional product, roadmap, and release updates from Airchor.
Subscription is separate from application access. You can unsubscribe from future updates.