OpenAI Log Image Renderer (Tampermonkey)
Render conversation images inline in OpenAI Platform conversation logs (/logs/conv_*), including:
input_image items with file_id- Markdown image links in assistant text
- Annotated-image placeholder flows (
[ANNOTATED_IMAGE])
This userscript is built for https://platform.openai.com/*.
GitHub
https://github.com/vonernue/openai-log-image-renderer
Contributing
Feel free to open an issue for bugs, questions, or feature requests:
https://github.com/vonernue/openai-log-image-renderer/issues
Pull requests are welcome:
https://github.com/vonernue/openai-log-image-renderer/pulls
Why This Script
In conversation logs, image-bearing items may appear as text or metadata instead of visible inline images.
This script intercepts the conversation items payload, resolves image URLs, and injects image cards directly into the matching message blocks.
Features
- Inline rendering inside the correct response/message card in the main logs UI
- Supports
file_id image resolution through:
- Reuses auth context from platform requests:
AuthorizationOpenAI-OrganizationOpenAI-Project
- Uses per-conversation header mapping from that conversation's
.../dashboard/conversations/{conv_id}/items request
- Retry UI for failed image loads
- Request dedupe + cooldown to reduce repeated failed calls / rate-limit pressure
- Automatic handling for SPA route changes
Installation
Option 1: Greasy Fork
- Install Tampermonkey in your browser.
- Open the Greasy Fork page for this script.
- Click Install this script.
Option 2: GitHub (manual)
- Install Tampermonkey.
- Open the script file from GitHub or the raw install URL above.
- Copy contents into a new Tampermonkey script, or install directly from the raw URL.
- Save and enable the script.
Usage
- Open
https://platform.openai.com/logs/conv_...
- Wait for the conversation items request to load.
- Images should appear inline under their corresponding message blocks.
If images do not appear immediately:
- refresh the page once
- ensure you have permission to view the files in that org/project
- click the script's Retry button on error badges
Configuration
Edit CONFIG at the top of the userscript:
UI.maxImageWidthPxUI.borderRadiusPxUI.showCaptionFEATURE_FLAGS.renderMarkdownImagesFEATURE_FLAGS.renderInputImageByFileIdFEATURE_FLAGS.renderAnnotatedImagePlaceholderDEBUG.enabled
Notes / Limitations
- This script depends on current OpenAI Platform DOM and internal API behavior; UI/API changes may require updates.
- Signed download URLs are time-limited.
- The script is intended for your own account/workspace usage where you are authorized to access those files.
- Not affiliated with OpenAI.
Privacy & Security
- Runs locally in your browser via Tampermonkey.
- Does not send data to third-party services.
- Uses existing authenticated platform context to request image download links.
Development
Repository structure:
userscripts/openai-log-image-renderer.user.jsREADME.mdREADME.github.md
To test quickly:
- Enable
DEBUG.enabled = true
- Reload the logs page
- Check browser console for
[OCI] logs
License
MIT
