Home

AXIOM AI Wiki

AXIOM — Build AI. Own AI.

A local-first toolkit for inspecting models and datasets, understanding available hardware, validating workspace readiness, and drafting AI project plans before committing compute.

Current documentation target: v0.3.0-beta.4

AXIOM combines a terminal-first CLI, a local Control Center, a loopback-first API, and workspace diagnostics. This Wiki documents the features that exist today and calls out boundaries where a design, estimate, or discovery result is not the same thing as execution.

Start here

Guide	Use it for
Getting Started	Install AXIOM, create a workspace, and open the Control Center.
What's New in beta.4	Module Explorer, workspace diagnostics, roadmap audit, support bundles, API Playground, and workflow design changes.
Control Center and Module Explorer	Understand the GUI sections and what they actually do.
CLI Guide	Find the right command for common project tasks.
Workspace Diagnostics and Support Bundles	Run checks and understand what a support archive contains.
Local API and Security	Daemon routes, loopback defaults, API Playground guardrails, and trust boundaries.
Workflow Design and Execution Boundaries	Distinguish saved designs and dry runs from actual training or inference.
Troubleshooting	Collect useful debugging information without sharing secrets.

Beta.4 at a glance

• Discoverability: a searchable Module Explorer maps current GUI pages, CLI command families, and the live OpenAPI surface.
• Workspace health: axiom doctor and axiom doctor --json provide read-only diagnostic checks.
• Roadmap integrity: axiom roadmap audit --json audits coverage and status labels in the scoped 120-item roadmap; it does not rewrite the roadmap.
• Support: axiom support-bundle --output ./axiom-support-bundle.zip produces a minimized diagnostic archive for user review.
• Safer API exploration: the API Playground uses a fixed GET-only allowlist and bounded response handling.
• Workflow design: templates, revision history, duplication, validation, and dry-run resource estimates help with planning before external execution.
• Control Center: navigation and error diagnostics are clearer, with remembered administrator-session behavior.

The AXIOM contract

AXIOM does	AXIOM does not claim
Inspect available local metadata and dataset structure.	That a model is fully compatible or a dataset is semantically good.
Estimate fit and create hardware-aware plans.	That estimates are benchmarks or guarantees.
Save, revise, validate, and estimate workflow designs.	That a workflow plan executes training, inference, or checkpoint management.
Score precomputed evaluation predictions.	That evaluation launches inference.
Discover local runtimes and tools.	That discovery means a runtime or account is authenticated and connected.
Provide local diagnostics and a minimized support archive.	That metadata is automatically anonymous in every context.

Release and downloads

v0.3.0-beta.4 is a prerelease. The release page is the source of truth for attached packages and checksums. Check the release workflow for build/publish status. Do not infer that every asset is available merely because a tag exists.

Quick command path

axiom version
axiom guide
axiom init my-ai
cd my-ai
axiom project validate
axiom config validate
axiom system info
axiom doctor --json
axiom roadmap audit --json

For install choices and first-run details, continue to Getting Started.

---

Project links: Repository · Issues · Releases · Changelog