Skip to main content
Public alphaLocal-first agent runtime

Autonomous development, kept accountable

Ship agentic changes without losing the cockpit.

Phonton gives coding agents a local control plane: plan the work, route focused repo context, verify the output, and hand back evidence a developer can trust.

Control
Local state
Models
BYOK
Handoff
Verified diff
phonton runlocal control plane
Verification gateready for review
$ phonton plan "add rate limiting to auth"
 scope       middleware, config, tests
 context     14 source handles
 provider    gemini / budget locked
 verify      cargo test -p auth

$ phonton review latest
 status      verified
 diff        ready for approval
Plan DAG4 steps
Local memoryenabled
ProviderBYOK
No Phonton model proxyRepo memory stays localPlan previews before editsVerification before review

Product loop

A control surface for the full coding loop.

Phonton is shaped around the moments that make AI coding trustworthy: scope, context, execution, verification, and review.

01

Plan in the open

The goal becomes a visible task graph with risk, dependencies, and expected verification before edits start.

02

Pack only what matters

Source handles, repo memory, and semantic context keep model calls focused on the actual change.

03

Run locally

Work routes through the developer machine with provider choice, sandbox boundaries, and inspectable state.

04

Prove the handoff

Checks and review payloads decide whether the output is ready, failed, or needs human direction.

Trust boundary

The developer machine remains the source of authority.

Phonton keeps the agent loop inspectable by making memory, provider routing, verification, and review artifacts explicit instead of hiding them behind a hosted proxy.

Local store

Configuration, memory, task history, and review payloads stay on disk.

BYOK routing

Use OpenAI, Anthropic, Gemini, OpenRouter, Ollama, or compatible endpoints.

Flight log

Plans, retries, approvals, rejections, and rollback context remain explicit.

Measured claims

Benchmarks report provider, commit, checks, spend, and review burden.

CLI first

The first product moment is a diagnostic, not a demo.

Start with `phonton doctor`, confirm the local environment, preview a plan, then let verification decide what is ready for review.

Audience

For developers who want acceleration with audit trails.

Senior engineers

Keep the terminal workflow and gain an agent loop that produces inspectable plans and diffs.

Platform teams

Fit autonomous coding into existing trust, CI, secrets, and repository policy.

OSS maintainers

Let agents help while public stewardship still depends on readable evidence.

Next step

Install the CLI. Run the doctor. Review the plan.