OpenAI Marketplace partners for engineering teams
Which developer tools sit in the first OpenAI Marketplace partner list, what committed spend means for engineers, and how to test Codex review against them first.
This research library uses AI-assisted source research and drafting. Linked sources support product claims; analysis and proposed exercises are our interpretation. Unless an article documents a test and its results, do not read it as a hands-on review or an independently verified benchmark.
OpenAI Marketplace lets eligible enterprise customers apply part of an existing OpenAI commitment toward approved partner software. OpenAI announced it at DevDay on 29 September 2026 with 32 launch partners, and several of them are tools engineering teams already buy, such as CodeRabbit, Greptile, Datadog and Replit. The OpenAI Marketplace page has the partner list, a Request to buy option and a waitlist for partners.
What the OpenAI Marketplace is
The Marketplace is a buying route. Nothing in the announcement puts partner tools inside Codex. Your company uses committed OpenAI spend to pay for software from approved partners.
Three facts are confirmed today. The offer is for eligible enterprise customers. It applies part of an existing OpenAI commitment, not new money. Enterprise customers can express interest now. OpenAI has not published a self-serve path, and individual Plus or Pro plans have no role in it.
Which OpenAI Marketplace partners matter to developers?
The published partner list is broad. It covers creative tools, customer experience, legal, cybersecurity and more. This table pulls out the names an engineering team is most likely to touch. The category column is our short summary of each product, not OpenAI's wording, except where noted.
| Partner | Category | Where it meets a Codex workflow |
|---|---|---|
| CodeRabbit | AI code review on pull requests | Overlaps with Codex /review and automatic reviews |
| Greptile | AI code review with codebase context | Same overlap, compare on the same pull requests |
| Factory | Coding agents | Alternative or addition to Codex for delegated tasks |
| Replit, Lovable | App building from prompts | Prototypes that may later move into your main repo |
| Datadog | Monitoring and observability | Production signals your agents and reviewers rely on |
| Baseten | Open-source models, per OpenAI's recap | Serving models next to your OpenAI usage |
| Hex | Data notebooks and analytics | Data work outside the code editor |
| Notion, Glean | Docs and workplace search | Context sources for specs and design documents |
| Palo Alto Networks, CrowdStrike | Cybersecurity, per OpenAI's recap | Security tooling around the code Codex writes |
| Figma | Design, listed as creative in OpenAI's recap | Design input for front-end work |
The list also includes Adobe, Basis, Decagon, DevRev, ElevenLabs, Harvey, Higgsfield, HubSpot, Hyperagent, Legora, Manus, Ramp, Rogo, Runway, Salesforce, ServiceNow and Sierra.
What committed spend on dev tools means for engineers
For a developer, the practical change is who approves the purchase. A code review tool that used to sit on a team card may now come through the same contract as your OpenAI usage. That usually means procurement and whoever owns the OpenAI commitment get a say.
The money is shared. Every dollar of commitment applied to a partner is a dollar that no longer covers API tokens, ChatGPT seats or Codex usage. If your team is close to its commitment already, a partner purchase can look free until the usage bill arrives.
It also changes the comparison you should run. Codex already ships review features on all plans. The CLI /review command offers presets to review against a base branch, uncommitted changes, a commit, or custom instructions. The ChatGPT desktop app has a Code Review plugin, and with the Codex connector and automatic reviews enabled, Codex can review a GitHub pull request before you open it. GitHub support is generally available, and GitLab merge request support is in preview.
So the real question for CodeRabbit or Greptile is narrow: does it catch issues that Codex review misses on your code, often enough to pay for it?
What OpenAI has not documented yet
Keep these open until your account team answers them in writing:
- Which products or plans of each partner qualify.
- How much of a commitment can go to partners, and whether any ratio applies.
- Whether an existing contract with a partner can move under the commitment.
- Which regions and contract types count as eligible.
- How partner spend shows up in OpenAI usage and billing reports.
None of these appear in the material OpenAI published today. Do not assume a discount either, because OpenAI does not mention one.
Test Codex review against a partner tool before you buy
Run this comparison in one repository before anyone signs. It takes one sprint and gives procurement evidence instead of a demo.
- Pick ten recently merged pull requests with known follow-up fixes or review comments.
- Check out the head commit of each pull request, run
/reviewin the Codex CLI, and choose Review against a base branch. - Run the partner tool on the same pull requests, using its trial if one is available.
- Record each finding in a shared sheet: tool, file, issue, and whether a human reviewer agrees it is real.
- Count real findings unique to each tool, and false positives per pull request.
Give Codex your team's review rules first. Put standing rules in AGENTS.md, and use the Review instructions setting in the desktop app for criteria and reporting format. A fair test gives both tools the same guidance. We use the same evidence-first review habit in our methodology.
If the partner tool finds real issues that Codex misses, you have a concrete case for the Marketplace request. If it does not, keep the commitment for usage.
Pick your ten pull requests today and run the first /review pass before your next procurement meeting.
Further reading
Related training topics
Review is one step in the methodology.
Related research
Codex mobile CLI: runs that survive review
A codex mobile cli pattern for running Codex CLI from anywhere: verification latches, model pins, and connector rosters that keep remote runs reviewable.
Codex workflows: governance that lives in the repo
How to govern codex workflows from the repo: a connector roster, a ten-line done checklist, a slash catalog, and a verification latch reviewers can replay.
Fast mode is not the default: when fast models earn it
The fast model is a tradeoff you make on purpose: scope ledgers, replay sandwiches, and connector cards that keep fast agent runs reviewable.
Continue through the research archive
Newer research
OpenAI dots for developers: bug reports to tested PRs
A hands-on guide to ChatGPT dots for developer work: turning Slack bug reports and user feedback into tested pull requests through Codex, with the controls to set first.
Earlier research
OpenAI Private Intelligence: set up Private Safety Processing
How Zero Data Retention with Private Safety Processing works in the OpenAI API, how an admin turns it on per project, and what it covers when Codex sends code.