Compare AI coding tools and spec frameworks
Every comparison on this site in one place. Most pages describe the other product from its own docs or repository and say when that was checked. SpecWeave and SpecWithCode are our products, so read the pages with that in mind.
Spec-driven frameworks
Tools that give an AI coding agent a written spec to work from.
- Spec-driven development tools compared: OpenSpec, GitHub Spec Kit, the BMAD Method, Kiro and SpecWeave in one table.
- OpenSpec vs SpecWeave: a lasting record of system behaviour versus task evidence and handoff.
- GitHub Spec Kit vs SpecWeave: slash commands and plan files versus one spec and a task ledger.
- BMAD Method vs SpecWeave: agent roles and planning documents versus one short spec.
- Spec as code vs spec-driven and intent-first: what the three terms mean and how they differ.
Coding agents
How the agents themselves differ, and how to use more than one.
- Claude Code vs Codex: instruction files, skills, sessions and usage limits.
- AGENTS.md vs CLAUDE.md vs skills: which agent reads which file, and what belongs in a skill instead.
Desktop workspaces
Apps that run several coding agents for you.
- SpecWithCode vs OpenCode, T3 Code and cloud agents: what runs the agent, where code runs, subscriptions and price.
- SpecWithCode vs BridgeMind: price, platforms and where the record of the work is kept.
Switching guides
Moving a half-finished task from one agent to another.
- Claude Code handoff: continue in any agent: the full reference for
specweave handoffandspecweave pickup. - What goes in handoff.md: the file the next agent reads, with a template.
- Switch from Claude Code to Codex mid-task
- Switch from Claude Code to OpenCode mid-task
- Switch from Claude Code to Cursor mid-task
- Claude Code usage limit reached: every option when a limit hits.
- Claude Code handoff to Codex, recorded: a real run as a video tutorial.