I have searched for any existing and/or related issues.
I have searched for any existing and/or related discussions.
I have also searched in the CLOSED issues AND CLOSED discussions and found no related items (your issue might already be addressed on the development branch!).
I am using the latest version of Open WebUI.
Installation Method
Other
Open WebUI Version
v0.8.10
Ollama Version (if applicable)
No response
Operating System
Windows 11
Browser (if applicable)
Vivaldi @ Latest
Confirmation
I have read and followed all instructions in README.md.
I am using the latest version of both Open WebUI and Ollama.
I have included the browser console logs.
I have included the Docker container logs.
I have provided every relevant configuration, setting, and environment variable used in my setup.
I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
Start with the initial platform/version/OS and dependencies used,
Specify exact install/launch/configure commands,
List URLs visited, user input (incl. example values/emails/passwords if needed),
Describe all options and toggles enabled or changed,
Include any files or environmental changes,
Identify the expected and actual result at each stage,
Ensure any reasonably skilled user can follow and hit the same issue.
Expected Behavior
Native mode enabled with modern models should be able to perform and complete code executions in chromium based browsers (including Vivaldi) and within the PWA.
This feature should allow models to complete calculations and validate assumptions.
Actual Behavior
Models are unable to complete the execution of the execute_code tool call. The entire model response either hangs or the execute_code call hangs with incomplete code/code execution. Web search tool executions and other tool executions work as expected as well.
Steps to Reproduce
Open vivaldi (have not tested on other chromium based browsers)
Perform a prompt that will lead a model to perform a calculation or run code
I do not have this issue on IOS PWA. Code executions work as expected.
Originally created by @jndao on GitHub (Mar 16, 2026).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/22720
### Check Existing Issues
- [x] I have searched for any existing and/or related issues.
- [x] I have searched for any existing and/or related discussions.
- [x] I have also searched in the CLOSED issues AND CLOSED discussions and found no related items (your issue might already be addressed on the development branch!).
- [x] I am using the latest version of Open WebUI.
### Installation Method
Other
### Open WebUI Version
v0.8.10
### Ollama Version (if applicable)
_No response_
### Operating System
Windows 11
### Browser (if applicable)
Vivaldi @ Latest
### Confirmation
- [x] I have read and followed all instructions in `README.md`.
- [x] I am using the latest version of **both** Open WebUI and Ollama.
- [x] I have included the browser console logs.
- [x] I have included the Docker container logs.
- [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.**
- [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
- [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps:
- Start with the initial platform/version/OS and dependencies used,
- Specify exact install/launch/configure commands,
- List URLs visited, user input (incl. example values/emails/passwords if needed),
- Describe all options and toggles enabled or changed,
- Include any files or environmental changes,
- Identify the expected and actual result at each stage,
- Ensure any reasonably skilled user can follow and hit the same issue.
### Expected Behavior
Native mode enabled with modern models should be able to perform and complete code executions in chromium based browsers (including Vivaldi) and within the PWA.
This feature should allow models to complete calculations and validate assumptions.
### Actual Behavior
Models are unable to complete the execution of the `execute_code` tool call. The entire model response either hangs or the `execute_code` call hangs with incomplete code/code execution. Web search tool executions and other tool executions work as expected as well.
### Steps to Reproduce
1. Open vivaldi (have not tested on other chromium based browsers)
2. Perform a prompt that will lead a model to perform a calculation or run code
3. Observe hanging `execute_code` process.
### Logs & Screenshots
<img width="539" height="66" alt="Image" src="https://github.com/user-attachments/assets/81012319-0431-40ef-880b-2e7f578fce2b" />
### Additional Information
This might be related to https://github.com/pyodide/pyodide/issues/5927 or https://github.com/open-webui/open-webui/discussions/9720
I do _not_ have this issue on IOS PWA. Code executions work as expected.
GiteaMirror
added the bug label 2026-05-18 05:16:57 -05:00
GLM
Today at 11:05 PM
Executing execute_code...
Input
code
# Extended test for background tab throttling validation
# This test runs longer to give time to switch tabs
import time
import sys
import os
print("=" * 60)
print("PYODIDE BACKGROUND TAB THROTTLING VALIDATION TEST")
print("=" * 60)
print("\n📁 Checking persistent file system...")
upload_dir = '/mnt/uploads'
if os.path.exists(upload_dir):
files = os.listdir(upload_dir)
print(f" Upload directory: {files if files else '(empty)'}")
else:
print(" Upload directory not found (fresh session)")
print("\n⏱️ Running timed operations...")
print(" (Switch to another tab NOW to test background throttling)")
print()
# Record timestamps to detect throttling
timestamps = []
for i in range(10):
start_iter = time.time()
# Do some computation
result = sum([x**2 for x in range(10000)])
elapsed = time.time() - start_iter
timestamps.append(elapsed)
print(f" Iteration {i+1:2d}: {elapsed:.4f}s (cumulative: {sum(timestamps):.2f}s)")
# Small sleep to simulate realistic workload
time.sleep(0.2)
print("\n" + "=" * 60)
# Analysis
total_time = sum(timestamps)
avg_time = total_time / len(timestamps)
expected_time = 0.2 * 10 # minimum expected
print("RESULTS:")
print(f" Total execution time: {total_time:.2f}s")
print(f" Average iteration: {avg_time:.4f}s")
print(f" Expected minimum: ~{expected_time:.1f}s (10 iterations × 0.2s)")
# Check for significant throttling (iterations taking >1s would indicate background throttling)
throttled = any(t > 1.0 for t in timestamps)
if throttled:
print("\n⚠️ THROTTLING DETECTED - Some iterations took >1 second")
print(" This indicates background tab throttling occurred")
else:
print("\n✅ No significant throttling detected")
print(" Execution ran smoothly without background delays")
print("=" * 60)
# Basic validation
print(f"\n🐍 Python version: {sys.version}")
print(f"📊 Math test: 2+2 = {2+2}")
print("\n✅ TEST COMPLETE")
Continues to hang. It seems to occur when the tab/PWA and the chat is not in focus. Not clear if this is an issue that can be worked around.
<!-- gh-comment-id:4067192419 -->
@jndao commented on GitHub (Mar 16, 2026):
Also ran in PWA on windows AND on firefox
```
GLM
Today at 11:05 PM
Executing execute_code...
Input
code
# Extended test for background tab throttling validation
# This test runs longer to give time to switch tabs
import time
import sys
import os
print("=" * 60)
print("PYODIDE BACKGROUND TAB THROTTLING VALIDATION TEST")
print("=" * 60)
print("\n📁 Checking persistent file system...")
upload_dir = '/mnt/uploads'
if os.path.exists(upload_dir):
files = os.listdir(upload_dir)
print(f" Upload directory: {files if files else '(empty)'}")
else:
print(" Upload directory not found (fresh session)")
print("\n⏱️ Running timed operations...")
print(" (Switch to another tab NOW to test background throttling)")
print()
# Record timestamps to detect throttling
timestamps = []
for i in range(10):
start_iter = time.time()
# Do some computation
result = sum([x**2 for x in range(10000)])
elapsed = time.time() - start_iter
timestamps.append(elapsed)
print(f" Iteration {i+1:2d}: {elapsed:.4f}s (cumulative: {sum(timestamps):.2f}s)")
# Small sleep to simulate realistic workload
time.sleep(0.2)
print("\n" + "=" * 60)
# Analysis
total_time = sum(timestamps)
avg_time = total_time / len(timestamps)
expected_time = 0.2 * 10 # minimum expected
print("RESULTS:")
print(f" Total execution time: {total_time:.2f}s")
print(f" Average iteration: {avg_time:.4f}s")
print(f" Expected minimum: ~{expected_time:.1f}s (10 iterations × 0.2s)")
# Check for significant throttling (iterations taking >1s would indicate background throttling)
throttled = any(t > 1.0 for t in timestamps)
if throttled:
print("\n⚠️ THROTTLING DETECTED - Some iterations took >1 second")
print(" This indicates background tab throttling occurred")
else:
print("\n✅ No significant throttling detected")
print(" Execution ran smoothly without background delays")
print("=" * 60)
# Basic validation
print(f"\n🐍 Python version: {sys.version}")
print(f"📊 Math test: 2+2 = {2+2}")
print("\n✅ TEST COMPLETE")
```
Continues to hang. It seems to occur when the tab/PWA and the chat is not in focus. Not clear if this is an issue that can be worked around.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @jndao on GitHub (Mar 16, 2026).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/22720
Check Existing Issues
Installation Method
Other
Open WebUI Version
v0.8.10
Ollama Version (if applicable)
No response
Operating System
Windows 11
Browser (if applicable)
Vivaldi @ Latest
Confirmation
README.md.Expected Behavior
Native mode enabled with modern models should be able to perform and complete code executions in chromium based browsers (including Vivaldi) and within the PWA.
This feature should allow models to complete calculations and validate assumptions.
Actual Behavior
Models are unable to complete the execution of the
execute_codetool call. The entire model response either hangs or theexecute_codecall hangs with incomplete code/code execution. Web search tool executions and other tool executions work as expected as well.Steps to Reproduce
execute_codeprocess.Logs & Screenshots
Additional Information
This might be related to https://github.com/pyodide/pyodide/issues/5927 or https://github.com/open-webui/open-webui/discussions/9720
I do not have this issue on IOS PWA. Code executions work as expected.
@jndao commented on GitHub (Mar 16, 2026):
Also ran in PWA on windows AND on firefox
Continues to hang. It seems to occur when the tab/PWA and the chat is not in focus. Not clear if this is an issue that can be worked around.
@tjbck commented on GitHub (Mar 22, 2026):
Not supported.