Issue Station
Sign In
Tools hereToolbox

hermes - 💡(How to fix) Fix [Bug]: Stream stalled mid tool-call [3 comments, 3 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
NousResearch/hermes-agent#18019•Fetched 2026-05-01 05:54:19
View on GitHub
Comments
3
Participants
3
Timeline
6
Reactions
0
Author
hosonfunghosonfung
Participants
alt-glitchalt-glitchhosonfunghosonfungmichaelmanlymichaelmanly
Timeline (top)
commented ×3labeled ×3

Error Message

  1. Hermes cannot create diagram and prompt error

Additional Logs / Traceback (optional)

Root Cause

Root Cause Analysis (optional)

Code Example

Debug report saved locally. Here's a summary of what it contains:

    Environment:
    - Version: 0.11.0 (2026.4.23) [9a145406]
    - OS: Linux 6.6.87.2-microsoft-standard-WSL2 x86_64
    - Python: 3.11.15, OpenAI SDK: 2.32.0
    - Profile: default, Hermes home: ~/.hermes

    Model/Provider:
    - Model: qwen35, Provider: custom (https://xxxxxx/v1/)

    API Keys: None of the standard provider keys are set (using custom provider)

    Features:
    - Toolsets: hermes-cli
    - MCP servers: 0
    - Memory: built-in
    - Gateway: running (systemd user, pid 5904)
    - Platforms: Telegram
    - Cron jobs: 0
    - Skills: 92

    Config overrides: display.streaming: True

    Notable issues from agent.log:
    - Several 404 errors for model qwen35 not found on the custom endpoint (sessions 20260501_001241_79e6a0, 48c3de26b9fc, 3fe820701396) — these happened around 00:19-00:23 when the model name may have been temporarily unavailable or misrouted
    - One API timeout (session c6a923c0cf7e at 00:34 — request timed out after 3 retries)
    - Vision tool errors: "Invalid image source" when an invalid URL/path was provided (22:50)
    - Context length detection failing for qwen35 — defaulting to 256,000 tokens
    - Gateway polling conflicts early on (22:31) from multiple bot instances, but resolved after restart
    - One "Message is too long" Telegram approval failure at 23:15 (fell back to text approval)

    Gateway log: Clean startup at 22:33, Telegram connected and stable since. No user allowlist configured (warning present).

    The full debug report was generated and displayed inline — nothing was uploaded externally since you used --local

---
RAW_BUFFERClick to expand / collapse

Bug Description

Hermes still meeting Stream stalled mid tool-call issure when it write larger file. Already fix according to other issue, but still same problem

<img width="1330" height="668" alt="Image" src="https://github.com/user-attachments/assets/5d1ddd2e-e847-4d99-aef5-3a83beaf253d" />

Steps to Reproduce

  1. ask hermes to create a diagram about local AI meeting notes
  2. Hermes start to research
  3. Hermes start to create diagram
  4. Hermes crash

Expected Behavior

  1. ask hermes to create a diagram about local AI meeting notes
  2. Hermes start to research
  3. Hermes start to create diagram
  4. Herems create diagram successfully

Actual Behavior

  1. ask hermes to create a diagram about local AI meeting notes
  2. Hermes start to research
  3. Hermes start to create diagram
  4. Hermes cannot create diagram and prompt error

Affected Component

CLI (interactive chat)

Messaging Platform (if gateway-related)

No response

Debug Report

Debug report saved locally. Here's a summary of what it contains:

    Environment:
    - Version: 0.11.0 (2026.4.23) [9a145406]
    - OS: Linux 6.6.87.2-microsoft-standard-WSL2 x86_64
    - Python: 3.11.15, OpenAI SDK: 2.32.0
    - Profile: default, Hermes home: ~/.hermes

    Model/Provider:
    - Model: qwen35, Provider: custom (https://xxxxxx/v1/)

    API Keys: None of the standard provider keys are set (using custom provider)

    Features:
    - Toolsets: hermes-cli
    - MCP servers: 0
    - Memory: built-in
    - Gateway: running (systemd user, pid 5904)
    - Platforms: Telegram
    - Cron jobs: 0
    - Skills: 92

    Config overrides: display.streaming: True

    Notable issues from agent.log:
    - Several 404 errors for model qwen35 not found on the custom endpoint (sessions 20260501_001241_79e6a0, 48c3de26b9fc, 3fe820701396) — these happened around 00:19-00:23 when the model name may have been temporarily unavailable or misrouted
    - One API timeout (session c6a923c0cf7e at 00:34 — request timed out after 3 retries)
    - Vision tool errors: "Invalid image source" when an invalid URL/path was provided (22:50)
    - Context length detection failing for qwen35 — defaulting to 256,000 tokens
    - Gateway polling conflicts early on (22:31) from multiple bot instances, but resolved after restart
    - One "Message is too long" Telegram approval failure at 23:15 (fell back to text approval)

    Gateway log: Clean startup at 22:33, Telegram connected and stable since. No user allowlist configured (warning present).

    The full debug report was generated and displayed inline — nothing was uploaded externally since you used --local

Operating System

wsl in win11

Python Version

No response

Hermes Version

No response

Additional Logs / Traceback (optional)

Root Cause Analysis (optional)

No response

Proposed Fix (optional)

No response

Are you willing to submit a PR for this?

  • I'd like to fix this myself and submit a PR

extent analysis

TL;DR

The issue with Hermes stalling when writing larger files may be related to the custom model provider or API timeouts, and verifying the model endpoint and increasing timeout settings could help resolve the issue.

Guidance

  • Check the custom model provider endpoint (https://xxxxxx/v1/) to ensure it is correctly configured and accessible.
  • Review the agent.log for recurring 404 errors or API timeouts, and consider increasing the timeout settings or adjusting the retry mechanism.
  • Verify that the display.streaming config override is set to True and that the Telegram platform is properly configured.
  • Investigate the "Context length detection failing" issue for the qwen35 model, as it may be related to the stalling problem.

Example

No code snippet is provided as the issue is more related to configuration and model provider setup.

Notes

The issue may be specific to the custom model provider or the WSL environment, and further debugging may be required to identify the root cause.

Recommendation

Apply workaround: Adjust the timeout settings and verify the model endpoint to see if it resolves the issue, as the problem may be related to the custom provider or API connectivity.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#api#indexing error#inference speed#output truncation#response parsing

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

hermes - 💡(How to fix) Fix [Bug]: Stream stalled mid tool-call [3 comments, 3 participants]