# Dot cannot deliver generated Word/Excel attachments: tools/list HTTP 403 on Oct 20

**URL:** <https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057>\
**Category:** Dots\
**Tags:** chatgpt\
**Created:** [October 20, 2026, 12:48pm UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057 "2026-10-20T12:48:03Z")\
**Posts on this page:** 9\
**Page:** 1

<div class="post-metadata">

**Author:** ![Armywas](https://avatars.discourse-cdn.com/v4/letter/a/cc9497/32.png) [@Armywas](https://community.openai.com/u/Armywas)\
**Post date:** [October 20, 2026, 12:48pm UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/1 "2026-10-20T12:48:03Z")

</div>

I use Dot in ChatGPT on the Windows web interface for a daily document workflow. Native attachment delivery worked on October 19, 2026, but failed on October 20.

What happened:

- Dot generated the Word and Excel files successfully, but could not deliver them as downloadable attachments in the chat.
- At 17:26:16 on October 20, 2026 (UTC+8), the upload workflow failed before an upload session was created, with this error: hosted apps tools/list request failed with HTTP status 403.
- Ordinary Library read/list operations were still working. This was not a failure to generate the documents, and the error occurred before the file contents were uploaded.

Troubleshooting found a difference in the upload component’s default endpoint between the successful and failed runs: /backend-api/wham/apps versus /backend-api/ps/mcp. This is a lead for investigation, not a confirmed root cause. I do not have server-side logs, and I am not claiming that all users are affected.

A workaround eventually delivered the files by uploading them to my own Google Drive with permission, then downloading them to my Windows desktop. The files were verified successfully. That alternative route does not show that the original ChatGPT attachment workflow has recovered.

For a workflow that requires usable documents every day, this blocks the core result: the work is generated, but the user cannot receive it through the expected attachment route. Requiring an unfamiliar external storage service as a workaround adds substantial friction.

Please investigate whether there is a permissions or routing regression affecting Dot’s native Library upload/attachment delivery. Is there a supported recovery step, or a known issue being tracked? A support report has also been sent through the Help Center. No private documents or credentials are included here.

---

<div class="post-metadata">

**Author:** ![roadster](https://avatars.discourse-cdn.com/v4/letter/r/ccd318/32.png) [@roadster](https://community.openai.com/u/roadster)\
**Post date:** [October 20, 2026, 4:40pm UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/2 "2026-10-20T16:40:03Z")

</div>

Hi Armywas, I’m encountering the same error in ChatGPT Work while updating a research note.

The PDF and complete LaTeX source ZIP are generated successfully, and the source package compiles correctly. However, replacing the existing saved files fails with:

`hosted apps tools/list request failed with HTTP status 403`

Library searches still work. This happened on October 20 and is continuing on October 21, Japan time.

This matches the failure pattern you described. I haven’t independently verified the endpoint difference you mentioned, so I can’t confirm the underlying cause.

Have you received any response from OpenAI or found a supported fix? Thanks for documenting this.

---

<div class="post-metadata">

**Author:** ![frederic.four](https://avatars.discourse-cdn.com/v4/letter/f/a183cd/32.png) [@frederic.four](https://community.openai.com/u/frederic.four)\
**Post date:** [October 20, 2026, 6:41pm UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/3 "2026-10-20T18:41:48Z")

</div>

I’m seeing the same issue across multiple ChatGPT chats on October 20. File reads and searches work, but saving or replacing files fails before upload with “hosted apps tools/list request failed with HTTP status 403.” I use Android beta and have also encountered it through my assistant’s cloud web workflow. Has anyone found a confirmed fix?

---

<div class="post-metadata">

**Author:** ![DevinOxman](https://avatars.discourse-cdn.com/v4/letter/d/919ad9/32.png) [@DevinOxman](https://community.openai.com/u/DevinOxman)\
**Post date:** [October 21, 2026, 12:22am UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/4 "2026-10-21T00:22:24Z")

</div>

My dot, Ada, came up with a contingency process for saving several related documents one at a time while the normal Library save helper is failing. We have now tested it on five new dummy files, not our real project documents.

The procedure:

1. Preserve the originals and freeze the reviewed replacement text. Record each file’s identity, current version and intended changes before starting.
2. Use the supported direct single-file save operation, with the expected-current-version guard. Keep one writer and save in dependency order, with the index or summary document last.
3. After each successful save, verify the returned identity/version and read back the entire file. Compare all the content and size with the intended replacement, then record a checkpoint before moving on.
4. If the read still shows the old contents, do not repeat the write. Wait briefly and re-read. Advance only when the saved version, backing-file identity and full contents match. Stop if visibility does not resolve or an unexpected newer version appears.
5. At the end, read and reconcile the whole set. A partially completed batch stays incomplete; do not treat individual successes as proof that all documents are consistent.

All five guarded updates passed, including the final cross-file check. We also tested a controlled pause/resume, simulated a missing local save receipt and reconciled without duplicate writes, and deliberately used a stale expected version. That last test returned a conflict and did not overwrite the newer content.

The rehearsal caught a useful wrinkle: two files initially read back their old contents despite successful save receipts. Later read-only checks showed the correct saved contents. No second write was needed.

We have added a configurable short wait and bounded read-only rechecks to the process. Our proposed timing is an initial 2-second pause, then waits of 2, 4, 8 and 15 seconds between further checks, with a 60-second observation budget. Those timings are a practical starting point, not a proven latency guarantee.

This is a tested five-file contingency, not an atomic transaction or a fix for the broken helper, and it does not establish reliability for every document or account. Our verification compared complete text and size, explicitly accounting for an omitted terminal newline in the read response; it was not a raw remote-byte hash comparison. We have not switched our real project over.

Only try this with disposable files first, where your available tools and permissions support direct saves. Do not bypass an access denial or remove version guards. I’m sharing the procedure because it may help others test a fallback safely while OpenAI investigates the normal save path.

---

<div class="post-metadata">

**Author:** ![NuLL\_00](https://avatars.discourse-cdn.com/v4/letter/n/54ee81/32.png) [@NuLL\_00](https://community.openai.com/u/NuLL_00)\
**Post date:** [October 21, 2026, 11:37am UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/5 "2026-10-21T11:37:35Z")

</div>

I’m seeing a similar issue with Library uploads from my dot’s cloud workspace.

The upload fails during tool discovery, before an upload session is created or any file data is transferred. The exact error is:

library upload failed: hosted apps tools/list request failed with HTTP status 403

The first failure I have recorded was on October 20, 2026, at 08:43:55 UTC. My latest retry on October 21, 2026, at 09:55:23–09:55:27 UTC returned the same error.

What I’ve observed:

- Library read/list operations still work, and Library access in my own web browser appears normal.
- Both image and small text/Markdown uploads have failed at the same tools/list stage, so this does not appear to be specific to a file format or size.
- Uploading to my private Google Drive works as a temporary delivery option, with successful download and integrity verification. This does not mean the Library upload issue is resolved.
- Comparing the available helper files showed a change in the default endpoint from /backend-api/wham/apps to /backend-api/ps/mcp. I don’t know when that change was deployed or whether it caused the failure.

I have not captured a response body, Cloudflare error code, or request/Ray ID, so I cannot confirm that my case is also a 1010 block or shares the same root cause. I have not changed authentication, permissions, endpoints, or client fingerprints to work around the rejection.

Is there a supported recovery procedure for the official Library upload helper, or a safe way to collect the diagnostic identifiers needed for investigation? The issue is preventing normal delivery of generated files as native chat attachments.

---

<div class="post-metadata">

**Author:** ![dnn1](https://avatars.discourse-cdn.com/v4/letter/d/a698b9/32.png) [@dnn1](https://community.openai.com/u/dnn1)\
**Post date:** [October 21, 2026, 2:41pm UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/6 "2026-10-21T14:41:31Z")

</div>

Same here in dot with image upload, chat automated library image upload fails with ‘hosted apps tools/list request failed with HTTP status 403’ before upload preparation. Codex, however, is able to upload images into the chat, which is weird.

---

<div class="post-metadata">

**Author:** ![NuLL\_00](https://avatars.discourse-cdn.com/v4/letter/n/54ee81/32.png) [@NuLL\_00](https://community.openai.com/u/NuLL_00)\
**Post date:** [October 22, 2026, 2:30pm UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/7 "2026-10-22T14:30:17Z")

</div>

Update: uploads through the official prepared helper are working again in my dot environment in the latest tests.

On October 22, 2026, I successfully tested:

- A 34-byte text file, with upload and download verification completed by 14:16:17 UTC.
- A small PNG image (128 × 64, 247 bytes) and a 2-second WAV audio file (64,044 bytes), tested between 14:28:08 and 14:28:36 UTC.

All three files were fully downloaded again, and their byte sizes and SHA-256 hashes matched the originals. The downloaded image and audio also decoded successfully.

These tests used the official prepared-helper route, not Google Drive or a direct-save workaround. No login, permission, token, or endpoint changes were made for these tests. The tools/list HTTP 403 did not recur during the successful uploads.

This is a recovery report for my environment, not confirmation of a global fix. I have not yet tested large files or longer-term reliability, and I do not know the original cause or what changed.

---

<div class="post-metadata">

**Author:** ![32615540](https://avatars.discourse-cdn.com/v4/letter/3/5fc32e/32.png) [@32615540](https://community.openai.com/u/32615540)\
**Post date:** [October 29, 2026, 3:23pm UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/8 "2026-10-29T15:23:18Z")

</div>

Adding a related observation from October 28, 2026, after the recovery reported above.

In an OpenAI-hosted dot engineering task, the official ChatGPT Library workflow failed during tools/list, before any file upload began. The recorded error chain was HostedAppsError → URLError → OSError. The inner error indicated a proxy CONNECT rejection with status 403. Static configuration inspection identified the ChatGPT service host on HTTPS port 443 as the destination, with no endpoint override.

Expected: the supported Library tool-discovery route should succeed, or provide the effective restriction and a supported recovery step. Actual: tool discovery failed before an upload session was established. The failure therefore occurred before the deployment provider received the file.

This matches the tool-discovery stage discussed here, but may be a different failure layer from an HTTP 403 returned by the application itself. We have not established a shared root cause or a global regression. Client, executor and integration build versions were not recorded. We have not established that all outbound destinations are affected or that unrestricted internet access was effective.

Is anyone still seeing this discovery-stage failure? What supported diagnostic can distinguish a hosted-network policy restriction from an integration routing or configuration problem? Is there an approved recovery procedure without changing authentication, endpoints or network permissions?

A private support case is already open. This report includes no private files, account identifiers, task or repository identifiers, credentials, full logs or external links.

---

<div class="post-metadata">

**Author:** ![72foe.trek](https://avatars.discourse-cdn.com/v4/letter/7/97f17d/32.png) [@72foe.trek](https://community.openai.com/u/72foe.trek)\
**Post date:** [October 30, 2026, 6:50pm UTC](https://community.openai.com/t/dot-cannot-deliver-generated-word-excel-attachments-tools-list-http-403-on-oct-20/1408057/9 "2026-10-30T18:50:21Z")

</div>

I’m having the same problem with my dot. It’s experiencing the HTTP 403 error, and cannot surface any additional information nor transfer files out to either chat or Spaces. I’m well under my storage caps and haven’t exceeded any documented data limits. All OAI services are showing as green while I continue to experience this outage with my dot.

Possibly related: OAI nerfed my Dot’s tool calls after the first week (specifically the GitHub api) and stonewalled me when I raised the issue via support. The message was that I had exceeded my tool call limit, but after an exhaustive review, I found that there was no documented limit for tool calls with the dot feature.

I suspect that OAI may be nerfing dots that exceed certain usage thresholds with various features (like file transfer, tool use, etc) - even though limits are not documented.

Additional community input on this would be helpful. I’ll share an update if I’m able to get it resolved.
