# Atomic

> Atomic is the verifiable coding agent runtime. Developers build repeatable engineering loops and execute them as inspectable workflows with checks, artifacts, gates, checkpoints, review, and human input.

Atomic is the runtime. Loops are repeatable agent systems built and run on it. Workflows are executable TypeScript definitions that give those loops structure and controls.

Atomic can be used to build verifiable software factories: broader software-production systems composed from engineering loops with verification built into their workflows. This is a system architecture enabled by the runtime, not a claim that installing Atomic creates an autonomous factory or guarantees correct or secure results.

## Product

- [Atomic homepage](https://bastani.ai/): Product overview, capabilities, and installation
- [Verifiable coding agent runtime](https://bastani.ai/verifiable-coding-agent-runtime/): Runtime, workflow, loop, loop-engineering, and verifiable software-factory terminology
- [Coding agent workflows](https://bastani.ai/coding-agent-workflows/): Software engineering workflow design, dynamic branching, composition, stages, tools, and inspectable handoffs
- [Workflow SDK for TypeScript](https://bastani.ai/coding-agent-workflow-sdk/): Current authoring syntax, typed contracts, tasks, parallel branches, artifacts, human input, and composition
- [Atomic built-in workflow examples](https://bastani.ai/coding-agent-workflow-examples/): Selected implementation and design workflows, plus six reusable control-flow patterns with distinct orchestration and verification shapes
- [Verification for coding agents](https://bastani.ai/verification-for-coding-agents/): Checks, evidence, artifacts, gates, reviewers, checkpoints, and human intervention
- [Recursive state machines and coding agent execution](https://bastani.ai/recursive-state-machines-for-coding-agents/): Finite state machines, recursive state machines, recursive language models, and the acyclic execution graph Atomic materializes and validates
- [Atomic resources](https://bastani.ai/resources/): Editorial index of evergreen guides, maintained docs, source, releases, examples, and selected field notes
- [Case Study: GRiDD Technologies](https://bastani.ai/case-studies/gridd/): GRiDD Technologies' jira-ship workflow across 12 Jira tickets — 18 of 19 PRs merged with zero reverts, ~375 verification commands run inside the workflows, and one prevented production incident

## Documentation

- [Agent-readable documentation index](https://docs.bastani.ai/llms.txt): Maintained technical documentation index for agents
- [Workflow documentation](https://docs.bastani.ai/workflows): Workflow authoring, execution, controls, and durable operations
- [Atomic documentation](https://docs.bastani.ai/): Complete product documentation

## Source

- [Atomic GitHub repository](https://github.com/bastani-inc/atomic): Source code, README, releases, and issue tracker
- [Workflow package source](https://github.com/bastani-inc/atomic/tree/main/packages/workflows): Public workflow authoring and runtime implementation
- [Repository workflow definitions](https://github.com/bastani-inc/atomic/tree/main/.atomic/workflows): Substantial project workflows with deterministic gates