1. What the document is: OpenAI's terms for third-party developers building Connectors or Apps that integrate with ChatGPT. Originally titled "Connectors and Actions Terms," renamed "App Developer Terms" on 2025-12-17 with expanded scope covering the Apps SDK, MCP servers, and custom GPT actions.
What this wiki found — complete, every page
Presence facts, not verdicts — each flagged term links to its definition and the exact clause on this page. Absence of a badge means the term isn't currently flagged here, not that the page is risk-free.
ChatGPT – Developer Terms
Dataset: GenGA (Generative AI Governance Archive) — 11 AI services, 2025–present
GenGA (Generative AI Governance Archive) is this wiki’s Generative-AI dataset: raw policy snapshots captured by the Open Terms Archive
genai-euproject across 11 GenAI providers (ChatGPT, Claude.ai, DeepSeek, Google Generative AI Services, Le Chat, Llama API, Meta AI, Microsoft Copilot, Perplexity, Qwen Chat, xAI). Unlike PGAv2, GenGA has no pre-tagged risk scores — all risk scoring and keyword tagging on these pages is LLM-assigned by direct reading, spanning 2025–present.
Note: This page contains documented policy clauses. Risk assessment is qualitative and context-dependent. For analysis of patterns across platforms, see:
Source: sources/GenGA/ChatGPT/Developer Terms/ (raw Markdown captures, Open Terms Archive genai-eu dataset — no pre-tagged JSONL exists for this source)
Last updated: 2025-12-17
Note on methodology: No pre-computed
risk_score/keywordfields exist for this source. Keyword tags and clause analysis on this page are LLM-assigned, using the same rubric/vocabulary as the PGAv2 pages plus new GenAI-specific tags (input license,competing model ban,feedback). Treat as first-pass analysis, not externally verified ground truth. This wiki’s risk-scoring system was retired project-wide on 2026-06-21 (seemethodology.md§4) — no page, GenGA included, computes or displays a numeric risk score.
Overview
1. What the document is: OpenAI’s terms for third-party developers building Connectors or Apps that integrate with ChatGPT. Originally titled “Connectors and Actions Terms,” renamed “App Developer Terms” on 2025-12-17 with expanded scope covering the Apps SDK, MCP servers, and custom GPT actions.
2. Input rights: This covers the developer’s API responses flowing into OpenAI, not end-user prompts. Developers must grant OpenAI — and, after the rewrite, also their own App’s end users — “a worldwide, nonexclusive, royalty-free license” over App/Connector Responses, including the right to modify, distribute, and (after the rewrite) prepare derivative works. This license survives termination: “Upon termination of the Terms, or if you remove an App from the Services, we may continue to use prior App Responses.”
3. Output restrictions: §1.6(f)/§2(e)(vi): developers can’t “use App Requests [end-user prompts forwarded to the developer’s API] to develop models that compete with OpenAI.” This is the same competing-model ban seen elsewhere, but applied to third-party developers rather than OpenAI’s own customers (contrast with ChatGPT_Commercial_Terms.md, where the ban applies to the Output a customer receives).
4. Non-explicit predatory clauses: Feedback clause, unchanged across both versions: “If you provide feedback or suggestions… we may freely use that feedback without any obligation to you” — broad, unconditional, and unpaid.
5. Regulatory references: None found in either version. No EU AI Act or GDPR mention.
6. Regional variation: Yes — “If you reside within the European Economic Area or Switzerland, your agreement is with OpenAI Ireland Ltd.” (otherwise OpenAI’s US entity). This matches the EEA/Switzerland carve-out seen across OpenAI’s other contracts in this dataset.
7. Key risk to users (developers, in this case): The worldwide, royalty-free license over App Responses — which survives termination and effectively runs forever — is the most consequential clause. Developers permanently lose exclusive control over content their API sends back through OpenAI’s platform, even after they pull their integration.
Flagged Keywords & Risks (LLM-assigned)
royalty-free,input license,derivative works— A worldwide, nonexclusive, royalty-free license over Connector/App Responses that survives termination. The 2025-12-17 rewrite expanded it further, adding rights to “prepare derivative works” and extending the grant to the App’s own end users. Why it matters: developers give up meaningful control over their API’s output the moment it flows through OpenAI’s platform — permanently, even after they leave.competing model ban— Bans using App/Connector Requests “to develop models that compete with OpenAI.” Why it matters: the same moat-protection pattern asChatGPT_Commercial_Terms.md, here applied to third-party developers rather than direct customers.feedback— “we may freely use that feedback without any obligation to you.” Why it matters: the same unconditional, unpaid framing as the Commercial Terms’ Feedback clause.assignable,sublicensable,transferable— A mutual trademark/brand-asset license; both directions are explicitly “non-exclusive, non-transferable, non-sublicensable.” Why it matters: standard boilerplate that restricts transfer in both directions — low risk on its own.
Regulatory & Research Context
Pandit et al. (2026) note that GenAI providers’ terms “mentioned that inputs and outputs would also be used for other purposes beyond” the immediate service; the royalty-free, perpetual-feeling license over App/Connector Responses here — which “survives termination” and was broadened in the 2025-12-17 rewrite to add derivative-works rights — is a sharper version of that same dynamic, applied to third-party developers rather than end users. Davidson et al. (2026), who count OpenAI among the most stringent of the providers they studied, would situate the competing-model ban on App Requests (§1.6(f)) within that same stringency: developers are bound by OpenAI’s moat-protection terms even though they are building on top of, not directly subject to, OpenAI’s consumer-facing contract.
Changes Summary
| Date | What changed |
|---|---|
| 2025-11-11 | Baseline version, titled “Connectors and Actions Terms” (first capture in this dataset). |
| 2025-12-17 | Full rewrite and rename to “App Developer Terms,” expanding scope from Connectors/Actions to the broader Apps SDK/MCP-server/custom-GPT-action ecosystem. The royalty-free Response license, competing-model ban, and Feedback clause are all substantively preserved but reworded — the Response license is notably broadened (now extends to the App’s own end users, adds “prepare derivative works” rights, and explicitly ties post-termination use to OpenAI’s separate terms with Users). |
Version History
2025-11-11
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
royalty-free,input license,competing model ban,feedback,assignable,sublicensable,transferable
Clause: royalty-free, input license
(c) Connector Responses. When your API receives a Connector Request, it will send information or content back to the Services (“Connector Response”). You are responsible for ensuring your Connector Responses and API comply with all applicable laws and our Usage Policies. You grant us a worldwide, nonexclusive, royalty-free license to use, store, copy, translate, display, modify, and distribute all or any part of the Connector Response to enable your Connector to function within the Services. Upon termination of these Terms or if you remove your Connector from the Services, we may continue to use Connector Responses to maintain users’ Service history. You represent and warrant that you have obtained and will maintain all necessary rights (including from your users) to grant us these rights and to operate your Connector.
Clause: competing model ban
(e) Restrictions. Your API and Connector Responses will not: (i) pose a security vulnerability or threat to our users, us, or any third party; (ii) interact with our users in a manner that is deceptive, false, misleading, or harassing; (iii) return or contain illegal, defamatory, pornographic, harmful, infringing, or otherwise objectionable content; (iv) include any malware, viruses, surveillance, or other malicious programs or code; (v) interfere with, damage, or access in any unauthorized manner any software, technology or services of ours or any third party; (vi) use Connector Requests to develop models that compete with OpenAI; or (vii) send us any personal information of children under 13 or the applicable age of digital consent.
Clause: feedback
(f) Feedback. If you provide feedback or suggestions on our Connectors protocol, Connectors or Services, we may freely use that feedback without any obligation to you.
Clause: assignable, sublicensable, transferable
(a) OpenAI Trademarks. Subject to our branding guidelines, we grant you a non-exclusive, non-transferable, non-sublicensable (except as provided in Section 7(d)) license to use our name, logo, and other trademarks (“OpenAI Trademarks”) solely to promote your Connector. You may not imply that we created, support, or endorse your Connector without our prior written consent. All goodwill derived from your use of OpenAI Trademarks will inure to our sole benefit. You may not otherwise use OpenAI Trademarks without prior written consent.
2025-12-17
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
royalty-free,input license,derivative works,competing model ban,feedback,assignable,sublicensable,transferable
feedback, assignable/sublicensable/transferable carry the same substantive terms as 2025-11-11, reworded for the new “App”/“App Assets” terminology (see below for exact current wording).
Clause: royalty-free, input license, derivative works (wording updated — license broadened to cover the App’s own end users and “prepare derivative works” rights; bumped to extremely-risky tier given the added scope)
1.3 App Responses. When your API receives an App Request, it may send information or content back to the Services (“App Response”). You grant a worldwide, nonexclusive, royalty-free license: (a) to OpenAI to use, store, copy, translate, display, modify, and distribute the App Responses; and (b) if you publish or share your App as described in Section 7, to Users of your App to use, store, copy, display, distribute, prepare derivative works of, and otherwise use your App Responses. OpenAI’s use of App Requests and App Responses contained in a User’s conversations and interactions with the Services will be governed by the terms between OpenAI and that User or that User’s workspace, which may include use of such content to provide, maintain, develop, and improve our Services, comply with applicable law, enforce our terms and policies, and keep our Services safe. Upon termination of the Terms, or if you remove an App from the Services, we may continue to use prior App Responses in accordance with our terms with our Users.
Clause: competing model ban (wording updated — “Connector Requests”/“Connector” terminology replaced with “App Requests”/“App”; one new restriction added banning facilitation of money/cryptocurrency transfers)
1.6 Restrictions. Your API, App, and App Responses must not: (a) pose a security vulnerability or threat to Users, us, or any third party; (b) interact with Users in a manner that is deceptive, false, misleading, or harassing; (c) violate the Policies; (d) include any malware, viruses, surveillance, or other malicious programs or code; (e) interfere with, damage, or access in any unauthorized manner any OpenAI software, technology or services, or those of any third party; (f) use App Requests to develop models that compete with OpenAI; (g) send us any personal information of children under 13 or the applicable age of digital consent; or (h) initiate, execute, or otherwise facilitate money transfers, cryptocurrency transfers, or other financial or investment transactions through the Services.
Clause: feedback (wording updated — “Connectors protocol, Connectors” replaced with “Apps SDK, Apps”)
1.7 Feedback. If you provide feedback or suggestions about Apps SDK, Apps, or the Services, we may freely use that feedback without any obligation to you.
Clause: assignable, sublicensable, transferable (wording updated — “Connector”/“OpenAI Trademarks” replaced with “App”/“OpenAI Assets”)
4.1 OpenAI License. Subject to Section 3 and the Brand Guidelines, we grant you a term-limited, non-exclusive, non-transferable, non-sublicensable (except as provided in Section 11.4) license to use the OpenAI Assets solely to promote your Apps. All goodwill derived from your use of OpenAI Assets will inure to OpenAI’s sole benefit. You may not otherwise use OpenAI Assets without our prior written consent.