# API

**URL:** https://community.openai.com/c/api/7.md

[Latest](https://community.openai.com/latest.md) · [Categories](https://community.openai.com/categories.md) · [Tags](https://community.openai.com/tags.md)

---

## [Read this before posting a new question](https://community.openai.com/t/read-this-before-posting-a-new-question/51)

<div class="topic-metadata">

**Author:** [@system](https://community.openai.com/u/system)\
**Replies:** 2\
**Last updated:** [February 21, 2023, 11:46am UTC](https://community.openai.com/t/read-this-before-posting-a-new-question/51 "2023-02-21T11:46:27Z")

</div>

Welcome to the OpenAI Developer Community, a forum for developers, users, and community members to meet, learn, share, and help one another while using and building with OpenAI products and developer tools. Looking for …

---

## [About the OpenAI API category](https://community.openai.com/t/about-the-openai-api-category/35)

<div class="topic-metadata">

**Author:** [@logankilpatrick](https://community.openai.com/u/logankilpatrick)\
**Replies:** 0\
**Last updated:** [March 10, 2021, 8:39pm UTC](https://community.openai.com/t/about-the-openai-api-category/35 "2021-03-10T20:39:25Z")

</div>

Questions, feedback, and best practices around building with OpenAI’s API. Please read the API docs before posting. ChatGPT questions, bug reports, feedback, feature requests, use cases, and troubleshooting are welcome…

---

## [Gpt-image-2 image creator and editor app - Python, desktop UI (download in latest \_J reply!)](https://community.openai.com/t/gpt-image-2-image-creator-and-editor-app-python-desktop-ui-download-in-latest-j-reply/1370104)

<div class="topic-metadata">

**Author:** [@\_j](https://community.openai.com/u/_j)\
**Replies:** 22\
**Last updated:** [October 25, 2026, 10:24am UTC](https://community.openai.com/t/gpt-image-2-image-creator-and-editor-app-python-desktop-ui-download-in-latest-j-reply/1370104 "2026-10-25T10:24:46Z")

</div>

GPT-Image Desktop Editor: a desktop GUI for OpenAI image generation and vision-guided edits A Christmas present by me for you - an image playground beyond OpenAI’s offerings. MOD NOTE: Download the latest version in the…

---

## [Short background noises (transients) interrupting Realtime voice agents and breaking the conversation, how are you handling this?](https://community.openai.com/t/short-background-noises-transients-interrupting-realtime-voice-agents-and-breaking-the-conversation-how-are-you-handling-this/1408109)

<div class="topic-metadata">

**Author:** [@lnobrega](https://community.openai.com/u/lnobrega)\
**Replies:** 1\
**Last updated:** [October 24, 2026, 2:37pm UTC](https://community.openai.com/t/short-background-noises-transients-interrupting-realtime-voice-agents-and-breaking-the-conversation-how-are-you-handling-this/1408109 "2026-10-24T14:37:24Z")

</div>

Hi everyone, We run voice agents in production on the Realtime API (gpt-realtime-mini, calls in Portuguese) and we’re seeing a recurring issue: short noises like a cough, a door, a keyboard or background chatter trigger…

---

## [OCR for Hebrew Text Images](https://community.openai.com/t/ocr-for-hebrew-text-images/1409232)

<div class="topic-metadata">

**Author:** [@misokol](https://community.openai.com/u/misokol)\
**Replies:** 1\
**Last updated:** [October 24, 2026, 9:00am UTC](https://community.openai.com/t/ocr-for-hebrew-text-images/1409232 "2026-10-24T09:00:31Z")

</div>

I am trying to use an Open AI model to process Biblical Hebrew Text. I have success with Chat GPT providing the image file and a text file of skills but using the API and a number of models does not completely work. Mo…

---

## [GPT 5.6 Don't want to Work?](https://community.openai.com/t/gpt-5-6-dont-want-to-work/1395684)

<div class="topic-metadata">

**Author:** [@arm200022](https://community.openai.com/u/arm200022)\
**Replies:** 1\
**Last updated:** [October 23, 2026, 9:39pm UTC](https://community.openai.com/t/gpt-5-6-dont-want-to-work/1395684 "2026-10-23T21:39:43Z")

</div>

The 5.6 Tier model doesn’t fully complete the task—it can stop and simply not continue, but I’d say that while it performs the action, it essentially does nothing but waste tokens, nerves, and so on! Then when you say “C…

---

## [Introducing Scheduled Automations to the Agents API](https://community.openai.com/t/introducing-scheduled-automations-to-the-agents-api/1409223)

<div class="topic-metadata">

**Author:** [@VeitB](https://community.openai.com/u/VeitB)\
**Replies:** 0\
**Last updated:** [October 23, 2026, 9:24pm UTC](https://community.openai.com/t/introducing-scheduled-automations-to-the-agents-api/1409223 "2026-10-23T21:24:34Z")

</div>

You can now schedule a message for an existing session or have it create a new one. Use a cron expression for recurring tasks or a timestamp to run something once in the future. Agents can also manage their own sche…

---

## [What is the default level of reasoning for Gpt 6 Astra](https://community.openai.com/t/what-is-the-default-level-of-reasoning-for-gpt-6-astra/1408343)

<div class="topic-metadata">

**Author:** [@wadesolowoniuk](https://community.openai.com/u/wadesolowoniuk)\
**Replies:** 6\
**Last updated:** [October 23, 2026, 7:27pm UTC](https://community.openai.com/t/what-is-the-default-level-of-reasoning-for-gpt-6-astra/1408343 "2026-10-23T19:27:49Z")

</div>

What’s Astra’s default reasoning level? It isn’t listed in the developer docs; only GPT-6.1 Sol appears.

---

## [Account credits gone event though I was just told I had credit](https://community.openai.com/t/account-credits-gone-event-though-i-was-just-told-i-had-credit/1396301)

<div class="topic-metadata">

**Author:** [@oases.plusher-0j](https://community.openai.com/u/oases.plusher-0j)\
**Replies:** 3\
**Last updated:** [October 23, 2026, 7:21pm UTC](https://community.openai.com/t/account-credits-gone-event-though-i-was-just-told-i-had-credit/1396301 "2026-10-23T19:21:54Z")

</div>

I have not used the API for a while and therefor had some credit left in the account, but today the balance is at zero and I have to add more to get my prompts working. I also just received an email saying I still had cr…

---

## [How reliable are enforced API hard spending limits under high concurrency?](https://community.openai.com/t/how-reliable-are-enforced-api-hard-spending-limits-under-high-concurrency/1408782)

<div class="topic-metadata">

**Author:** [@devzaib](https://community.openai.com/u/devzaib)\
**Replies:** 2\
**Last updated:** [October 24, 2026, 8:55pm UTC](https://community.openai.com/t/how-reliable-are-enforced-api-hard-spending-limits-under-high-concurrency/1408782 "2026-10-24T20:55:36Z")

</div>

I’m trying to understand how OpenAI’s enforced hard spending limits behave when thousands of API requests are submitted concurrently. The documentation mentions that spending can slightly exceed the configured limit whi…

---

## [/v1/responses with background: true does not respect the json\_schema (with strict: true), works when background:false](https://community.openai.com/t/v1-responses-with-background-true-does-not-respect-the-json-schema-with-strict-true-works-when-background-false/1394809)

<div class="topic-metadata">

**Author:** [@julie.thomas](https://community.openai.com/u/julie.thomas)\
**Replies:** 4\
**Last updated:** [October 23, 2026, 3:55pm UTC](https://community.openai.com/t/v1-responses-with-background-true-does-not-respect-the-json-schema-with-strict-true-works-when-background-false/1394809 "2026-10-23T15:55:06Z")

</div>

When calling the Responses API with a json\_schema, option strict: true and background: true, the json\_schema is not respected. Some keys are missing and the descriptions are not followed. If I use background: false, the…

---

## [Whisper hallucination - how to recognize and solve?](https://community.openai.com/t/whisper-hallucination-how-to-recognize-and-solve/218307)

<div class="topic-metadata">

**Author:** [@nikola1jankovic](https://community.openai.com/u/nikola1jankovic)\
**Replies:** 27\
**Last updated:** [October 23, 2026, 3:38pm UTC](https://community.openai.com/t/whisper-hallucination-how-to-recognize-and-solve/218307 "2026-10-23T15:38:14Z")

</div>

Ok, I am using Whisper API for some time now. It works very good for big languages and almost acceptable for small ones. However, occasionally it hallucinates and as part of the transcription, it sends back repeated wor…

---

## [GPT-5.4-NANO inventing words? or a bug in tokens send in response?](https://community.openai.com/t/gpt-5-4-nano-inventing-words-or-a-bug-in-tokens-send-in-response/1391796)

<div class="topic-metadata">

**Author:** [@VivacityDesign](https://community.openai.com/u/VivacityDesign)\
**Replies:** 1\
**Last updated:** [October 23, 2026, 2:46pm UTC](https://community.openai.com/t/gpt-5-4-nano-inventing-words-or-a-bug-in-tokens-send-in-response/1391796 "2026-10-23T14:46:29Z")

</div>

Since about a week gpt-5.4-nano is inventing new words or mixing tokens randomly, producing wild responses. It starts well at each prompt, then progressively goes astray. It looks like it has something to do with transl…

---

## [Python user agents coming with proper web bot auth signature headers](https://community.openai.com/t/python-user-agents-coming-with-proper-web-bot-auth-signature-headers/1409085)

<div class="topic-metadata">

**Author:** [@1992sura](https://community.openai.com/u/1992sura)\
**Replies:** 1\
**Last updated:** [October 23, 2026, 1:27pm UTC](https://community.openai.com/t/python-user-agents-coming-with-proper-web-bot-auth-signature-headers/1409085 "2026-10-23T13:27:28Z")

</div>

In server logs, we observed requests labeled ChatGPT-Agent with Signature and Signature-Input headers, alg="ed25519", and tag="web-bot-auth". Some requests used Python-urllib/3.12 as their User-Agent. Our analysis showe…

---

## [Unexpected Refusal in GPT-6 Luna Decisions API During Content Classification](https://community.openai.com/t/unexpected-refusal-in-gpt-6-luna-decisions-api-during-content-classification/1406979)

<div class="topic-metadata">

**Author:** [@CinarCamur](https://community.openai.com/u/CinarCamur)\
**Replies:** 2\
**Last updated:** [October 23, 2026, 10:01am UTC](https://community.openai.com/t/unexpected-refusal-in-gpt-6-luna-decisions-api-during-content-classification/1406979 "2026-10-23T10:01:42Z")

</div>

Environment Surface: API (Python SDK) Endpoint or feature: Decisions API — client.decisions.create() Model: gpt-6-luna SDK and version: OpenAI Python SDK — \[version to be confirmed\] Bug I’m using the OpenAI…

---

## [Native Responses multi-agent mixes hosted Shell ownership and rejects a function result](https://community.openai.com/t/native-responses-multi-agent-mixes-hosted-shell-ownership-and-rejects-a-function-result/1408860)

<div class="topic-metadata">

**Author:** [@jhariani1](https://community.openai.com/u/jhariani1)\
**Replies:** 1\
**Last updated:** [October 23, 2026, 9:58am UTC](https://community.openai.com/t/native-responses-multi-agent-mixes-hosted-shell-ownership-and-rejects-a-function-result/1408860 "2026-10-23T09:58:52Z")

</div>

Reproduced using the official Python SDK 3.35.1, Python 3.12.3, gpt-6.1-sol, and the responses\_multi\_agent=v1 beta over WebSocket. The script below uses only one flat, non-deferred function and hosted Shell. It has no ap…

---

## [Results from toy experiments with Decisions API](https://community.openai.com/t/results-from-toy-experiments-with-decisions-api/1404010)

<div class="topic-metadata">

**Author:** [@platypus](https://community.openai.com/u/platypus)\
**Replies:** 22\
**Last updated:** [October 23, 2026, 8:11am UTC](https://community.openai.com/t/results-from-toy-experiments-with-decisions-api/1404010 "2026-10-23T08:11:00Z")

</div>

When something new/cool comes out I usually give it some toy experiments to see if it holds up, before applying it on a real world task. I decided to put Decisions API against classic statistics experiments: (1) loaded c…

---

## [API Returns 424 Error for MCP Calls](https://community.openai.com/t/api-returns-424-error-for-mcp-calls/1353593)

<div class="topic-metadata">

**Author:** [@guytp](https://community.openai.com/u/guytp)\
**Replies:** 7\
**Last updated:** [October 23, 2026, 5:57am UTC](https://community.openai.com/t/api-returns-424-error-for-mcp-calls/1353593 "2026-10-23T05:57:55Z")

</div>

I’m having an issue with integrating an MCP server during testing. To try and simplify this I am at the point of the simplest schema and request for it and I’ve removed all our code and am just testing through the API p…

---

## [Can anyone help with free OpenAI tokens with Data Sharing?](https://community.openai.com/t/can-anyone-help-with-free-openai-tokens-with-data-sharing/1391606)

<div class="topic-metadata">

**Author:** [@shalinijha](https://community.openai.com/u/shalinijha)\
**Replies:** 2\
**Last updated:** [October 22, 2026, 8:15pm UTC](https://community.openai.com/t/can-anyone-help-with-free-openai-tokens-with-data-sharing/1391606 "2026-10-22T20:15:06Z")

</div>

. i want to enable free token for my open ai , how can i do that . can anyone help me doing this

---

## [GPT 5.6 Luna+Sol inference in Australia only](https://community.openai.com/t/gpt-5-6-luna-sol-inference-in-australia-only/1391608)

<div class="topic-metadata">

**Author:** [@nullbio](https://community.openai.com/u/nullbio)\
**Replies:** 3\
**Last updated:** [October 22, 2026, 6:40pm UTC](https://community.openai.com/t/gpt-5-6-luna-sol-inference-in-australia-only/1391608 "2026-10-22T18:40:03Z")

</div>

Is there any way to use Luna or Sol in Australia without my data leaving Australian data-servers? In Australia, many businesses have a data protection requirement that limits them to keeping the data within Australian IR…

---

## [Error Message - This tool call was blocked by OpenAI's safety checks](https://community.openai.com/t/error-message-this-tool-call-was-blocked-by-openais-safety-checks/1408852)

<div class="topic-metadata">

**Author:** [@Alan.Harbar](https://community.openai.com/u/Alan.Harbar)\
**Replies:** 2\
**Last updated:** [October 22, 2026, 4:49pm UTC](https://community.openai.com/t/error-message-this-tool-call-was-blocked-by-openais-safety-checks/1408852 "2026-10-22T16:49:34Z")

</div>

Environment Surface: API / SDK / Playground / Platform Endpoint or feature: Model: SDK and version, if applicable: Bug What happened, and what did you expect? Reproduction Minimal code, request, or steps to reproduce: …

---

## [Gpt-audio-mini replacement](https://community.openai.com/t/gpt-audio-mini-replacement/1408821)

<div class="topic-metadata">

**Author:** [@tom84](https://community.openai.com/u/tom84)\
**Replies:** 1\
**Last updated:** [October 22, 2026, 3:39pm UTC](https://community.openai.com/t/gpt-audio-mini-replacement/1408821 "2026-10-22T15:39:04Z")

</div>

gpt-audio-mini has been deprecated and is facing shutdown next year, but there is no equivalent at the price point - the gpt-audio-1.5 it suggests is over 3x more expensive. We’ve built products around these models and …

---

## [What would make the OpenAI CLI more useful?](https://community.openai.com/t/what-would-make-the-openai-cli-more-useful/1408690)

<div class="topic-metadata">

**Author:** [@vguvvala](https://community.openai.com/u/vguvvala)\
**Replies:** 7\
**Last updated:** [October 22, 2026, 3:18pm UTC](https://community.openai.com/t/what-would-make-the-openai-cli-more-useful/1408690 "2026-10-22T15:18:41Z")

</div>

Hi everyone! I’m working on the OpenAI CLI, which lets you use the OpenAI API from your terminal. I’d love to hear how you use it and what would make it better. Are there workflows that are difficult, features you’d lik…

---

## [Images API repeatedly flags innocent romantic kissing as sexual content](https://community.openai.com/t/images-api-repeatedly-flags-innocent-romantic-kissing-as-sexual-content/1408741)

<div class="topic-metadata">

**Author:** [@aoifebayloife](https://community.openai.com/u/aoifebayloife)\
**Replies:** 10\
**Last updated:** [October 22, 2026, 2:15pm UTC](https://community.openai.com/t/images-api-repeatedly-flags-innocent-romantic-kissing-as-sexual-content/1408741 "2026-10-22T14:15:38Z")

</div>

I’ve encountered repeated false-positive moderation rejections while trying to generate ordinary romantic anime artwork using OpenAI’s Images Playground and API. The requested scene featured two fully clothed fictional …

---

## [GPT-Live starts calls talking about a buffet in a strange voice?](https://community.openai.com/t/gpt-live-starts-calls-talking-about-a-buffet-in-a-strange-voice/1402575)

<div class="topic-metadata">

**Author:** [@JakeTheSnake](https://community.openai.com/u/JakeTheSnake)\
**Replies:** 11\
**Last updated:** [October 22, 2026, 12:38pm UTC](https://community.openai.com/t/gpt-live-starts-calls-talking-about-a-buffet-in-a-strange-voice/1402575 "2026-10-22T12:38:53Z")

</div>

Ive had 2/251 GPT-Live calls (Appliance Repair CSR Ops) that start with a massive artifact where the voice is super exasperated talking about a ‘flexible buffet’, and its weird it happened twice because the script/phrasi…

---

## [API administrators can now restrict service tiers at organization and project level](https://community.openai.com/t/api-administrators-can-now-restrict-service-tiers-at-organization-and-project-level/1408738)

<div class="topic-metadata">

**Author:** [@VeitB](https://community.openai.com/u/VeitB)\
**Replies:** 0\
**Last updated:** [October 22, 2026, 8:21am UTC](https://community.openai.com/t/api-administrators-can-now-restrict-service-tiers-at-organization-and-project-level/1408738 "2026-10-22T08:21:29Z")

</div>

These policies apply to default, Flex, and Priority processing. Fast is treated as Priority, and restrictions set at the organization level take precedence over project settings. Scale Tier is not covered. If a request …

---

## [Realtime API hosted MCP: listings/calls stall 2–10 s or fail before connecting, only for a .ai hostname](https://community.openai.com/t/realtime-api-hosted-mcp-listings-calls-stall-2-10-s-or-fail-before-connecting-only-for-a-ai-hostname/1407759)

<div class="topic-metadata">

**Author:** [@huayuqin](https://community.openai.com/u/huayuqin)\
**Replies:** 2\
**Last updated:** [October 22, 2026, 7:18am UTC](https://community.openai.com/t/realtime-api-hosted-mcp-listings-calls-stall-2-10-s-or-fail-before-connecting-only-for-a-ai-hostname/1407759 "2026-10-22T07:18:36Z")

</div>

We use the Realtime API (gpt-realtime-2.1, WebSocket) with a remote MCP server attached as a hosted tool (type “mcp”, require\_approval “never”). Since 14 Oct 2026, tool listings and calls are intermittently delayed by 2–…

---

## [Migrating MCP servers for an existing app](https://community.openai.com/t/migrating-mcp-servers-for-an-existing-app/1390167)

<div class="topic-metadata">

**Author:** [@anikets](https://community.openai.com/u/anikets)\
**Replies:** 1\
**Last updated:** [October 21, 2026, 5:27pm UTC](https://community.openai.com/t/migrating-mcp-servers-for-an-existing-app/1390167 "2026-10-21T17:27:13Z")

</div>

We have an existing ChatGPT app. We are trying to migrate from a 3P managed MCP server to an in house MCP server. As part of this migration, we need to publish a new app version with a new base URL. The UI prevents us fr…

---

## [Feature Request: Add a Configurable Loudness Parameter to the Voice API](https://community.openai.com/t/feature-request-add-a-configurable-loudness-parameter-to-the-voice-api/1408528)

<div class="topic-metadata">

**Author:** [@botpulse](https://community.openai.com/u/botpulse)\
**Replies:** 0\
**Last updated:** [October 21, 2026, 5:06pm UTC](https://community.openai.com/t/feature-request-add-a-configurable-loudness-parameter-to-the-voice-api/1408528 "2026-10-21T17:06:05Z")

</div>

I’m developing AI voice agents for production customer-service phone calls and would like to request a configurable loudness control in the Voice API. The Create Voice API does not currently expose a dedicated parameter…

---

## [Consecutive MCP tool calls with UI widgets cause the previous widget to be replaced](https://community.openai.com/t/consecutive-mcp-tool-calls-with-ui-widgets-cause-the-previous-widget-to-be-replaced/1388538)

<div class="topic-metadata">

**Author:** [@cqy](https://community.openai.com/u/cqy)\
**Replies:** 3\
**Last updated:** [October 21, 2026, 4:17pm UTC](https://community.openai.com/t/consecutive-mcp-tool-calls-with-ui-widgets-cause-the-previous-widget-to-be-replaced/1388538 "2026-10-21T16:17:43Z")

</div>

Hi, I found a reproducible issue when building an MCP server with UI widgets attached to multiple tools. Problem When two MCP tools are called consecutively within the same assistant turn, and both tools have their own …

[Next page](https://community.openai.com/c/api/7.md?page=1)
