Theo Claw vs Codex
OpenAI's software engineering agent: a cloud sandbox per task, plus a CLI and an IDE extension
Codex is the best kind of narrow: give it a repository and a task and it works until there is a pull request to read. Theo Claw Code does the same job as one mode of a colleague that also watches an inbox, answers the phone, runs every morning, and asks you a question when it genuinely needs one.
Theo Claw is OpenCharts' always-on autonomous worker and its code mode is directly comparable to Codex: both clone a real repository into a cloud workspace, read and edit files, run commands and tests, and open a pull request you review. They diverge on scope and governance. The same Theo Claw worker can also hold a phone number, email address and calendar, listen in Slack, run named jobs on a schedule, pause to ask a plain question mid-task, earn autonomy one kind of action at a time, write a receipt for every run, and publish a finished run as a read-only page. Codex leads on terminal and editor integration, parallel coding tasks, and review tooling built around diffs.
Where each one wins
- The same worker also answers email, texts, calls and Slack, so shipping the fix and telling the customer are one job
- A choice of engines from several labs per worker, with cost tiers, or your own provider key
- A structured question mid-task, with one-tap options, instead of guessing or stopping
- Autonomy earned one kind of action at a time, with deleting never earnable
- Receipts for every run: the engine, pages, messages, teammates and credits
- A public read-only page for a finished run, with no repository access required
- Several named standing jobs per worker, each on its own schedule
- Work outside code lands as editable decks, documents, sheets and apps
- Terminal and editor integration built around an engineer's working day
- Starting many independent coding tasks and reviewing the results side by side
- Review workflows centered on diffs and existing pull requests
Theo Claw and Codex solve overlapping problems differently. The matrix below makes the trade-offs concrete, and the migration note covers what it takes to switch when you're ready.
The right pick for your team
Choose Theo Claw when…
- The job does not stop at the pull request: someone has to be emailed, called, or updated in Slack.
- You want one worker that ships code on Tuesday and clears an inbox every morning.
- Your team needs approvals, receipts, a trust ladder and org rules before an agent acts.
- You want to choose, or let Theo choose, which engine does the thinking and see what the run cost.
- You want to show somebody what the worker did without giving them repository access.
Choose Codex when…
- Your work is entirely code and your team lives in the terminal and editor.
- You want to start many independent coding tasks and review the diffs together.
- You want an agent that reviews existing pull requests as its main job.
Feature comparison
What it costs, what it takes
Pricing snapshot
Migrating from Codex
Frequently asked questions
Does Theo Claw work in my repository, or only scaffold code?
What does Theo Claw do that a coding agent does not?
Can a worker ask me a question instead of guessing?
Can I show somebody a run without giving them access?
Which one should a small engineering team pick?

Hire a worker that keeps working
Give a Theo Claw worker a goal, the tools it may use, and how much rope it gets. It runs while you sleep and asks before anything risky. Included from the Starter plan, with the whole workspace alongside it.
Included from Starter · Approvals and receipts on every run
