[GH-ISSUE #17626] issue: "Chunk too big" error when using Google Gemini 2.5 Flash with image input #57013

New Issue

GiteaMirror · 2026-05-05T20:24:48-05:00

GiteaMirror commented

2026-05-05 20:24:48 -05:00

Originally created by @BND-1 on GitHub (Sep 21, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/17626

Check Existing Issues

I have searched for any existing and/or related issues.
I have searched for any existing and/or related discussions.
I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

v0.6.30

Ollama Version (if applicable)

No response

Operating System

centos

Browser (if applicable)

chrome

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

When using the Google: Gemini 2.5 Flash Image Preview model, sending a prompt that includes an image results in a "Chunk too big" error in the user interface. This prevents the model from processing the request and providing a response.

Actual Behavior

The request fails, and a red error box appears in the chat interface with the message: Chunk too big.

Steps to Reproduce

1.Navigate to the OpenWebUI chat interface.
2.Select Google: Gemini 2.5 Flash Image Preview from the model selection dropdown.
3.Upload an image or provide an image along with a text prompt (e.g., "draw a puppy").
4.Send the message to the model.
5.Observe the error message displayed in the chat window.

Logs & Screenshots

Additional Information

This error suggests a potential issue with how large inputs, specifically image data, are chunked or sent to the backend/model API. It seems to be hitting a size limit somewhere in the data transmission pipeline. This issue currently blocks the use of the vision capabilities for the Gemini 2.5 Flash model within OpenWebUI.

Originally created by @BND-1 on GitHub (Sep 21, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/17626 ### 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 am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version v0.6.30 ### Ollama Version (if applicable) _No response_ ### Operating System centos ### Browser (if applicable) chrome ### 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 When using the Google: Gemini 2.5 Flash Image Preview model, sending a prompt that includes an image results in a "Chunk too big" error in the user interface. This prevents the model from processing the request and providing a response. ### Actual Behavior The request fails, and a red error box appears in the chat interface with the message: Chunk too big. ### Steps to Reproduce 1.Navigate to the OpenWebUI chat interface. 2.Select Google: Gemini 2.5 Flash Image Preview from the model selection dropdown. 3.Upload an image or provide an image along with a text prompt (e.g., "draw a puppy"). 4.Send the message to the model. 5.Observe the error message displayed in the chat window. ### Logs & Screenshots <img width="1945" height="449" alt="Image" src="https://github.com/user-attachments/assets/57b7fbb1-6da0-47ae-b0fc-643352a0820c" /> ### Additional Information This error suggests a potential issue with how large inputs, specifically image data, are chunked or sent to the backend/model API. It seems to be hitting a size limit somewhere in the data transmission pipeline. This issue currently blocks the use of the vision capabilities for the Gemini 2.5 Flash model within OpenWebUI.

GiteaMirror added the bug label 2026-05-05 20:24:48 -05:00

GiteaMirror closed this issue

2026-05-05 20:24:49 -05:00

GiteaMirror commented

2026-05-05 20:24:52 -05:00

@yborunov commented on GitHub (Sep 26, 2025):

same issue. Any update?

@yborunov commented on GitHub (Sep 26, 2025): same issue. Any update?

GiteaMirror commented

2026-05-05 20:24:56 -05:00

@cj1324 commented on GitHub (Sep 26, 2025):

same issue. Any update?

@cj1324 commented on GitHub (Sep 26, 2025): same issue. Any update?

GiteaMirror commented

2026-05-05 20:25:00 -05:00

@ShirasawaSama commented on GitHub (Sep 26, 2025):

We previously discovered that when SSE single-line data is extremely large (likely exceeding 16KB), directly using Python's asynchronous iterator to traverse aiohttp's response.content forces the use of aiohttp's built-in buffer size, which cannot be configured.

https://github.com/open-webui/open-webui/blob/main/backend/open_webui/routers/openai.py#L947

Our solution was to replace it with our own asynchronous iterator, calling AsyncContentStream's iter_chunks() method to iterate over all bytes. We actively split each line using the \n symbol and then yield the data for each line.

This approach avoids the constraints of aiohttp's buffer size. Unfortunately, I don't have permission to publish this specific code, so I can only share the underlying solution concept.

cc @tjbck

@ShirasawaSama commented on GitHub (Sep 26, 2025): We previously discovered that when SSE single-line data is extremely large (likely exceeding 16KB), directly using Python's asynchronous iterator to traverse aiohttp's `response.content` forces the use of aiohttp's built-in buffer size, which cannot be configured. https://github.com/open-webui/open-webui/blob/main/backend/open_webui/routers/openai.py#L947 Our solution was to replace it with our own asynchronous iterator, calling AsyncContentStream's `iter_chunks()` method to iterate over all bytes. We actively split each line using the `\n` symbol and then yield the data for each line. This approach avoids the constraints of aiohttp's buffer size. Unfortunately, I don't have permission to publish this specific code, so I can only share the underlying solution concept. cc @tjbck

GiteaMirror commented

2026-05-05 20:25:04 -05:00

@OakerO commented on GitHub (Sep 29, 2025):

Maybe someone should publish it then?

@OakerO commented on GitHub (Sep 29, 2025): Maybe someone should publish it then?

GiteaMirror commented

2026-05-05 20:25:06 -05:00

@rgaricano commented on GitHub (Sep 29, 2025):

approach, for 37d1c85c99/backend/open_webui/routers/openai.py (L929-L977)

Increased buffer & replaced the default async iterator on response.content uses line-based iteration, which buffers entire lines in memory (for SSE data exceeding 16KB on a single line, this forces the StreamReader to buffer the entire line until it finds a newline character).
By using iter_chunked() or iter_any(), we avoid this line-buffering behavior and process data in smaller, manageable chunks regardless of line boundaries.
We can choose iter_any() for immediate data availability or iter_chunked() with a specific chunk size (like 8192 bytes) for more predictable memory usage.

try:
    session = aiohttp.ClientSession(
        trust_env=True, 
        timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT),
        read_bufsize=128 * 1024  # Increased buffer size
    )

    r = await session.request(
        method="POST",
        url=request_url,
        data=payload,
        headers=headers,
        cookies=cookies,
        ssl=AIOHTTP_CLIENT_SESSION_SSL,
    )

    # Check if response is SSE
    if "text/event-stream" in r.headers.get("Content-Type", ""):
        streaming = True

        # Use chunk-based iteration instead of direct content
        async def sse_generator():
            async for chunk in r.content.iter_any():  # or iter_chunked(8192)
                yield chunk

        return StreamingResponse(
            sse_generator(),
            status_code=r.status,
            headers=dict(r.headers),
            background=BackgroundTask(
                cleanup_response, response=r, session=session
            ),
        )
# ...
# rest of code

@rgaricano commented on GitHub (Sep 29, 2025): approach, for https://github.com/open-webui/open-webui/blob/37d1c85c996e1bdcd505e1e6d62b2f17acd8df23/backend/open_webui/routers/openai.py#L929-L977 - Increased buffer & replaced the default async iterator on response.content uses line-based iteration, which buffers entire lines in memory (for SSE data exceeding 16KB on a single line, this forces the StreamReader to buffer the entire line until it finds a newline character). By using iter_chunked() or iter_any(), we avoid this line-buffering behavior and process data in smaller, manageable chunks regardless of line boundaries. We can choose iter_any() for immediate data availability or iter_chunked() with a specific chunk size (like 8192 bytes) for more predictable memory usage. ``` try: session = aiohttp.ClientSession( trust_env=True, timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT), read_bufsize=128 * 1024 # Increased buffer size ) r = await session.request( method="POST", url=request_url, data=payload, headers=headers, cookies=cookies, ssl=AIOHTTP_CLIENT_SESSION_SSL, ) # Check if response is SSE if "text/event-stream" in r.headers.get("Content-Type", ""): streaming = True # Use chunk-based iteration instead of direct content async def sse_generator(): async for chunk in r.content.iter_any(): # or iter_chunked(8192) yield chunk return StreamingResponse( sse_generator(), status_code=r.status, headers=dict(r.headers), background=BackgroundTask( cleanup_response, response=r, session=session ), ) # ... # rest of code ```

GiteaMirror commented

2026-05-05 20:25:07 -05:00

@ShirasawaSama commented on GitHub (Sep 29, 2025):

@rgaricano

In fact, many of my MCP SSE calls exceed 500KB, and occasionally even surpass 50MB per line. Consequently, my code includes logic to skip lines exceeding a certain length.

However, I don't think this is a common scenario, so it may not necessarily need to be implemented.

Additionally, we cannot directly call iter_any or iter_chunked and then yield the chunk directly; this approach is incorrect.

Because a single chunk may contain multiple lines, or a single line may be split across multiple chunks. Therefore, we must concatenate multiple chunks and use the \n character to separate lines of text. If the current chunk's terminator is not a \n character, we can assume that the next chunk or subsequent chunks still contain the remainder of the current line.

@ShirasawaSama commented on GitHub (Sep 29, 2025): @rgaricano In fact, many of my MCP SSE calls exceed 500KB, and occasionally even surpass 50MB per line. Consequently, my code includes logic to skip lines exceeding a certain length. However, I don't think this is a common scenario, so it may not necessarily need to be implemented. Additionally, we cannot directly call `iter_any` or `iter_chunked` and then yield the chunk directly; this approach is incorrect. Because a single chunk may contain multiple lines, or a single line may be split across multiple chunks. Therefore, we must concatenate multiple chunks and use the `\n` character to separate lines of text. If the current chunk's terminator is not a `\n` character, we can assume that the next chunk or subsequent chunks still contain the remainder of the current line.

GiteaMirror commented

2026-05-05 20:25:07 -05:00

@rgaricano commented on GitHub (Sep 29, 2025):

then maybe something like this:

try:
    session = aiohttp.ClientSession(
        trust_env=True, 
        timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT),
        read_bufsize=128 * 1024  # Increased buffer size
    )

    r = await session.request(
        method="POST",
        url=request_url,
        data=payload,
        headers=headers,
        cookies=cookies,
        ssl=AIOHTTP_CLIENT_SESSION_SSL,
    )

    # Check if response is SSE
    if "text/event-stream" in r.headers.get("Content-Type", ""):
        streaming = True
        buffer = b""

        async def sse_line_generator():
            """Generator that handles SSE line boundaries."""

            async for chunk in r.content.iter_any():
                buffer += chunk

                # Process complete lines
                while b"\n" in buffer:
                    line, buffer = buffer.split(b"\n", 1)
                    yield line + b"\n"

            # Handle any remaining incomplete line
            if buffer:
                yield buffer

        return StreamingResponse(
            sse_line_generator(),
            status_code=r.status,
            headers=dict(r.headers),
            background=BackgroundTask(
                cleanup_response, response=r, session=session
            ),
        )
# ...
# rest of  code

@rgaricano commented on GitHub (Sep 29, 2025): then maybe something like this: ``` try: session = aiohttp.ClientSession( trust_env=True, timeout=aiohttp.ClientTimeout(total=AIOHTTP_CLIENT_TIMEOUT), read_bufsize=128 * 1024 # Increased buffer size ) r = await session.request( method="POST", url=request_url, data=payload, headers=headers, cookies=cookies, ssl=AIOHTTP_CLIENT_SESSION_SSL, ) # Check if response is SSE if "text/event-stream" in r.headers.get("Content-Type", ""): streaming = True buffer = b"" async def sse_line_generator(): """Generator that handles SSE line boundaries.""" async for chunk in r.content.iter_any(): buffer += chunk # Process complete lines while b"\n" in buffer: line, buffer = buffer.split(b"\n", 1) yield line + b"\n" # Handle any remaining incomplete line if buffer: yield buffer return StreamingResponse( sse_line_generator(), status_code=r.status, headers=dict(r.headers), background=BackgroundTask( cleanup_response, response=r, session=session ), ) # ... # rest of code ```

GiteaMirror commented

2026-05-05 20:25:08 -05:00

@ddddavid-he commented on GitHub (Oct 8, 2025):

Any update?

@ddddavid-he commented on GitHub (Oct 8, 2025): Any update?

GiteaMirror commented

2026-05-05 20:25:09 -05:00

@ShirasawaSama commented on GitHub (Oct 11, 2025):

Any update? 🤔

@ShirasawaSama commented on GitHub (Oct 11, 2025): Any update? 🤔

GiteaMirror commented

2026-05-05 20:25:10 -05:00

@COCONUTwwater commented on GitHub (Oct 17, 2025):

Any update? 🤔

@COCONUTwwater commented on GitHub (Oct 17, 2025): Any update? 🤔

GiteaMirror commented

2026-05-05 20:25:11 -05:00

@Classic298 commented on GitHub (Oct 17, 2025):

PR welcome afaik

@Classic298 commented on GitHub (Oct 17, 2025): PR welcome afaik

GiteaMirror commented

2026-05-05 20:25:11 -05:00

@GeoffreyLng commented on GitHub (Oct 21, 2025):

Same issue, any updates ?

@GeoffreyLng commented on GitHub (Oct 21, 2025): Same issue, any updates ?

GiteaMirror commented

2026-05-05 20:25:12 -05:00

@Classic298 commented on GitHub (Oct 21, 2025):

No update yet, PR welcome

@Classic298 commented on GitHub (Oct 21, 2025): No update yet, PR welcome

GiteaMirror commented

2026-05-05 20:25:13 -05:00

@mweichert commented on GitHub (Oct 25, 2025):

Also experiencing this issue with gemini-2.5-flash-lite.

@mweichert commented on GitHub (Oct 25, 2025): Also experiencing this issue with `gemini-2.5-flash-lite`.

GiteaMirror commented

2026-05-05 20:25:14 -05:00

@NopeNix commented on GitHub (Oct 30, 2025):

same here

@NopeNix commented on GitHub (Oct 30, 2025): same here

GiteaMirror commented

2026-05-05 20:25:15 -05:00

@zuogod commented on GitHub (Nov 3, 2025):

Any update? 🤔

@zuogod commented on GitHub (Nov 3, 2025): Any update? 🤔

GiteaMirror commented

2026-05-05 20:25:16 -05:00

@Classic298 commented on GitHub (Nov 3, 2025):

PR welcome

@Classic298 commented on GitHub (Nov 3, 2025): PR welcome

GiteaMirror commented

2026-05-05 20:25:18 -05:00

@ShirasawaSama commented on GitHub (Nov 3, 2025):

#18884

@ShirasawaSama commented on GitHub (Nov 3, 2025): #18884

GiteaMirror commented

2026-05-05 20:25:20 -05:00

@silentoplayz commented on GitHub (Nov 5, 2025):

Related - https://github.com/open-webui/open-webui/issues/16935

@silentoplayz commented on GitHub (Nov 5, 2025): Related - https://github.com/open-webui/open-webui/issues/16935

GiteaMirror commented

2026-05-05 20:25:21 -05:00

@Classic298 commented on GitHub (Nov 6, 2025):

should be fixed in dev

@Classic298 commented on GitHub (Nov 6, 2025): should be fixed in dev

GiteaMirror commented

2026-05-05 20:25:23 -05:00

@ShirasawaSama commented on GitHub (Nov 7, 2025):

@Classic298 Not yet resolved

@ShirasawaSama commented on GitHub (Nov 7, 2025): @Classic298 Not yet resolved <img width="1078" height="890" alt="Image" src="https://github.com/user-attachments/assets/eb3e4405-72f1-440b-8367-1022fec5a8ae" />

GiteaMirror commented

2026-05-05 20:25:26 -05:00

@ddddavid-he commented on GitHub (Nov 7, 2025):

Same situation on my machine

@ddddavid-he commented on GitHub (Nov 7, 2025): Same situation on my machine

GiteaMirror commented

2026-05-05 20:25:28 -05:00

@DeusMaximus commented on GitHub (Nov 7, 2025):

Still happening through LiteLLM

@DeusMaximus commented on GitHub (Nov 7, 2025): Still happening through LiteLLM <img width="833" height="273" alt="Image" src="https://github.com/user-attachments/assets/614fad83-1b22-483b-96bd-7b8904960648" />

GiteaMirror commented

2026-05-05 20:25:32 -05:00

@GeoffreyLng commented on GitHub (Nov 7, 2025):

Still occuring when using LiteLLM

@GeoffreyLng commented on GitHub (Nov 7, 2025): Still occuring when using LiteLLM <img width="1081" height="486" alt="Image" src="https://github.com/user-attachments/assets/ba011592-e10c-4ebc-9e9a-a05ca717238b" />

GiteaMirror commented

2026-05-05 20:25:36 -05:00

@ddkedr commented on GitHub (Nov 7, 2025):

@Classic298 can we reopen it? Still does not work when using LiteLLM

@ddkedr commented on GitHub (Nov 7, 2025): @Classic298 can we reopen it? Still does not work when using LiteLLM

GiteaMirror commented

2026-05-05 20:25:39 -05:00

@Classic298 commented on GitHub (Nov 7, 2025):

I am aware, but that is as it looks to me, an issue with LiteLLM. I raised it with them, hopefully they can address it soon.

(Add Gemini 2.5 Flash Image in the Image configs)

@Classic298 commented on GitHub (Nov 7, 2025): I am aware, but that is as it looks to me, an issue with LiteLLM. I raised it with them, hopefully they can address it soon. (Add Gemini 2.5 Flash Image in the Image configs)

GiteaMirror commented

2026-05-05 20:25:40 -05:00

@AlexZorzi commented on GitHub (Nov 7, 2025):

@Classic298 for me its not working either, im not using LiteLLM, connected directly to OpenRouter.

@AlexZorzi commented on GitHub (Nov 7, 2025): @Classic298 for me its not working either, im not using LiteLLM, connected directly to OpenRouter. <img width="1047" height="271" alt="Image" src="https://github.com/user-attachments/assets/d015d547-f452-43dc-a217-b6b29880e789" />

GiteaMirror commented

2026-05-05 20:25:41 -05:00

@Classic298 commented on GitHub (Nov 7, 2025):

Do not talk to the model - add it in the image config as I said. - and not sure openrouter is supported yet, but the official Gemini API is supported.

@Classic298 commented on GitHub (Nov 7, 2025): Do not talk to the model - add it in the image config as I said. - and not sure openrouter is supported yet, but the official Gemini API is supported.

GiteaMirror commented

2026-05-05 20:25:42 -05:00

@ShirasawaSama commented on GitHub (Nov 10, 2025):

Hey everyone, the “Chunk too big” issue has been fixed in the latest dev branch.

You can enable it using the following environment variable:

CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760

And the issue where models in OpenRouter cannot output images will be fixed in this pull request #19073.

@ShirasawaSama commented on GitHub (Nov 10, 2025): Hey everyone, the “Chunk too big” issue has been fixed in the latest dev branch. You can enable it using the following environment variable: ```bash CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 ``` And the issue where models in OpenRouter cannot output images will be fixed in this pull request #19073.

GiteaMirror commented

2026-05-05 20:25:44 -05:00

@tjbck commented on GitHub (Nov 10, 2025):

@silentoplayz @Classic298 Documentation updated wanted here!

@tjbck commented on GitHub (Nov 10, 2025): @silentoplayz @Classic298 Documentation updated wanted here!

GiteaMirror commented

2026-05-05 20:25:45 -05:00

@silentoplayz commented on GitHub (Nov 10, 2025):

You can enable it using the following environment variable:
CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760

What is enabled when I set this as an environment variable for my Open WebUI instance? Is it necessary to have this environment variable set at all for image generation/editing to work properly using Google Gemini 2.5 Flash in Open WebUI?

@silentoplayz commented on GitHub (Nov 10, 2025): > You can enable it using the following environment variable: > `CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760` What is enabled when I set this as an environment variable for my Open WebUI instance? Is it necessary to have this environment variable set at all for image generation/editing to work properly using Google Gemini 2.5 Flash in Open WebUI?

GiteaMirror commented

2026-05-05 20:25:46 -05:00

@ShirasawaSama commented on GitHub (Nov 10, 2025):

Related PR #19076

@ShirasawaSama commented on GitHub (Nov 10, 2025): Related PR #19076

GiteaMirror commented

2026-05-05 20:25:46 -05:00

@ShirasawaSama commented on GitHub (Nov 10, 2025):

You can enable it using the following environment variable:
CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760

What is enabled when I set this as an environment variable for my Open WebUI instance? Is it necessary to have this environment variable set at all for image generation/editing to work properly using Google Gemini 2.5 Flash in Open WebUI?

This value is recommended to be greater than 5MB. As a reference, a GPT image typically outputs around 3.5MB of Base64-encoded image data. Our production environment is configured to 8MB (since we implemented sending socket.io data via SSE, which prevents massive single-line data from causing Redis cluster crashes; therefore, I do not recommend setting this value too high).

A value greater than 2.5MB must be set to display images generated by the nano banana model.

@ShirasawaSama commented on GitHub (Nov 10, 2025): > > You can enable it using the following environment variable: > > `CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760` > > What is enabled when I set this as an environment variable for my Open WebUI instance? Is it necessary to have this environment variable set at all for image generation/editing to work properly using Google Gemini 2.5 Flash in Open WebUI? This value is recommended to be greater than 5MB. As a reference, a GPT image typically outputs around 3.5MB of Base64-encoded image data. Our production environment is configured to 8MB (since we implemented sending socket.io data via SSE, which prevents massive single-line data from causing Redis cluster crashes; therefore, I do not recommend setting this value too high). A value greater than 2.5MB must be set to display images generated by the nano banana model.

GiteaMirror commented

2026-05-05 20:25:47 -05:00

@Classic298 commented on GitHub (Nov 10, 2025):

documted

@Classic298 commented on GitHub (Nov 10, 2025): documted

GiteaMirror commented

2026-05-05 20:25:48 -05:00

@usef814 commented on GitHub (Nov 11, 2025):

Hey everyone, the “Chunk too big” issue has been fixed in the latest dev branch.

You can enable it using the following environment variable:

CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760
And the issue where models in OpenRouter cannot output images will be fixed in this pull request #19073.

This will be great, I have been having the same issue ("Chunk too big") between openwebui and openrouter when trying to generate an image. . In the meantime using the integration with comfyui, works ok but text just isn't great when compared to gpt-5 image or many others.

@usef814 commented on GitHub (Nov 11, 2025): > Hey everyone, the “Chunk too big” issue has been fixed in the latest dev branch. > > You can enable it using the following environment variable: > > CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 > And the issue where models in OpenRouter cannot output images will be fixed in this pull request [#19073](https://github.com/open-webui/open-webui/pull/19073). This will be great, I have been having the same issue ("Chunk too big") between openwebui and openrouter when trying to generate an image. . In the meantime using the integration with comfyui, works ok but text just isn't great when compared to gpt-5 image or many others.

GiteaMirror commented

2026-05-05 20:25:48 -05:00

@Dzenn00 commented on GitHub (Nov 11, 2025):

Hi, we used the specified variable, but we get this output without a picture. Most often, the model simply sends ` symbol

@Dzenn00 commented on GitHub (Nov 11, 2025): Hi, we used the specified variable, but we get this output without a picture. Most often, the model simply sends ` symbol <img width="1028" height="226" alt="Image" src="https://github.com/user-attachments/assets/732f0790-ce4c-47a6-87e9-b289c9e72c6d" />

GiteaMirror commented

2026-05-05 20:25:49 -05:00

@M0n7y5 commented on GitHub (Nov 11, 2025):

CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE
does not help at all, i tested nano banana and chatgpt 5 image, still chunk too big error

@M0n7y5 commented on GitHub (Nov 11, 2025): CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE does not help at all, i tested nano banana and chatgpt 5 image, still chunk too big error

GiteaMirror commented

2026-05-05 20:25:50 -05:00

@Classic298 commented on GitHub (Nov 11, 2025):

@M0n7y5 that env var doesn't exist yet.

It was just introduced on the dev branch and will be shipped in the next version

@Classic298 commented on GitHub (Nov 11, 2025): @M0n7y5 that env var doesn't exist yet. It was just introduced on the dev branch and will be shipped in the next version

GiteaMirror commented

2026-05-05 20:25:53 -05:00

@ShirasawaSama commented on GitHub (Nov 12, 2025):

@Dzenn00 Are you using OpenRouter?

An issue where models in OpenRouter cannot output images will be fixed in this pull request #19073.

@ShirasawaSama commented on GitHub (Nov 12, 2025): @Dzenn00 Are you using OpenRouter? > An issue where models in OpenRouter cannot output images will be fixed in this pull request #19073.

GiteaMirror commented

2026-05-05 20:25:56 -05:00

@Dzenn00 commented on GitHub (Nov 12, 2025):

@ShirasawaSama yes

@Dzenn00 commented on GitHub (Nov 12, 2025): @ShirasawaSama yes

GiteaMirror commented

2026-05-05 20:25:59 -05:00

@M0n7y5 commented on GitHub (Nov 12, 2025):

@M0n7y5 that env var doesn't exist yet.

It was just introduced on the dev branch and will be shipped in the next version

Ah i see, my bad then. I'll wait for the update.

@M0n7y5 commented on GitHub (Nov 12, 2025): > [@M0n7y5](https://github.com/M0n7y5) that env var doesn't exist yet. > > It was just introduced on the dev branch and will be shipped in the next version Ah i see, my bad then. I'll wait for the update.

GiteaMirror commented

2026-05-05 20:26:01 -05:00

@ShirasawaSama commented on GitHub (Nov 19, 2025):

Hello to all users of OpenRouter here. The dev branch now supports image output in the Nano Banana model format for OpenRouter. If you have some free time, please git checkout to the latest dev branch for testing. Thanks! #19073

If you encounter any issues, feel free to reply directly below and I'll make adjustments immediately. Of course, other third-party model providers are also welcome to test it!

Thank you all for your testing—it helps make the feature more stable!

@ShirasawaSama commented on GitHub (Nov 19, 2025): Hello to all users of OpenRouter here. The `dev` branch now supports image output in the Nano Banana model format for OpenRouter. If you have some free time, please git checkout to the latest dev branch for testing. Thanks! #19073 If you encounter any issues, feel free to reply directly below and I'll make adjustments immediately. Of course, other third-party model providers are also welcome to test it! Thank you all for your testing—it helps make the feature more stable!

GiteaMirror commented

2026-05-05 20:26:02 -05:00

@silentoplayz commented on GitHub (Nov 19, 2025):

Hello to all users of OpenRouter here. The dev branch now supports image output in the Nano Banana model format for OpenRouter. If you have some free time, please git checkout to the latest dev branch for testing. Thanks! #19073

If you encounter any issues, feel free to reply directly below and I'll make adjustments immediately. Of course, other third-party model providers are also welcome to test it!

Thank you all for your testing—it helps make the feature more stable!

Tested and working on the dev branch for me. Amazing work!

@silentoplayz commented on GitHub (Nov 19, 2025): > Hello to all users of OpenRouter here. The `dev` branch now supports image output in the Nano Banana model format for OpenRouter. If you have some free time, please git checkout to the latest dev branch for testing. Thanks! [#19073](https://github.com/open-webui/open-webui/pull/19073) > > If you encounter any issues, feel free to reply directly below and I'll make adjustments immediately. Of course, other third-party model providers are also welcome to test it! > > Thank you all for your testing—it helps make the feature more stable! Tested and working on the `dev` branch for me. Amazing work! <img width="2560" height="1275" alt="Image" src="https://github.com/user-attachments/assets/15f57410-ac54-4ab2-a376-4e0646ba1ac4" /> <img width="2560" height="1275" alt="Image" src="https://github.com/user-attachments/assets/d9f3c385-66dc-4a3f-96b6-c865276a572b" />

GiteaMirror commented

2026-05-05 20:26:03 -05:00

@zuli12-dev commented on GitHub (Nov 22, 2025):

i just tested with the latest docker dev image (ghcr.io/open-webui/open-webui:git-b32f781) and CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 but i still see the same error

@zuli12-dev commented on GitHub (Nov 22, 2025): i just tested with the latest docker dev image (ghcr.io/open-webui/open-webui:git-b32f781) and CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 but i still see the same error

GiteaMirror commented

2026-05-05 20:26:04 -05:00

@ShirasawaSama commented on GitHub (Nov 23, 2025):

i just tested with the latest docker dev image (ghcr.io/open-webui/open-webui:git-b32f781) and CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 but i still see the same error

Have you tried setting this value to a larger number, such as 100000000? Or could you tell me who your service provider is? (e.g., OpenRouter)
Additionally, what method did you use to set environment variables in Docker?

@ShirasawaSama commented on GitHub (Nov 23, 2025): > i just tested with the latest docker dev image (ghcr.io/open-webui/open-webui:git-b32f781) and CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 but i still see the same error Have you tried setting this value to a larger number, such as 100000000? Or could you tell me who your service provider is? (e.g., OpenRouter) Additionally, what method did you use to set environment variables in Docker?

GiteaMirror commented

2026-05-05 20:26:04 -05:00

@YorhaL commented on GitHub (Nov 23, 2025):

Use dev Docker + CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE + LiteLLM + Vertex Gemini 3 Pro Image

It works now. thx

@YorhaL commented on GitHub (Nov 23, 2025): Use `dev` Docker + CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE + LiteLLM + Vertex Gemini 3 Pro Image <img width="1085" height="724" alt="Image" src="https://github.com/user-attachments/assets/2dac5b35-241f-4b9b-879a-1a9fb639b21c" /> It works now. thx

GiteaMirror commented

2026-05-05 20:26:05 -05:00

@zuli12-dev commented on GitHub (Nov 23, 2025):

i just tested with the latest docker dev image (ghcr.io/open-webui/open-webui:git-b32f781) and CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 but i still see the same error

Have you tried setting this value to a larger number, such as 100000000? Or could you tell me who your service provider is? (e.g., OpenRouter) Additionally, what method did you use to set environment variables in Docker?

Yes i just tried both, it seems it was some CORS quirk, it generates the images just fine now, many thanks!

@zuli12-dev commented on GitHub (Nov 23, 2025): > > i just tested with the latest docker dev image (ghcr.io/open-webui/open-webui:git-b32f781) and CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 but i still see the same error > > Have you tried setting this value to a larger number, such as 100000000? Or could you tell me who your service provider is? (e.g., OpenRouter) Additionally, what method did you use to set environment variables in Docker? Yes i just tried both, it seems it was some CORS quirk, it generates the images just fine now, many thanks! <img width="719" height="1173" alt="Image" src="https://github.com/user-attachments/assets/57acc6a1-1805-48db-93cf-d2c1059aba12" />

GiteaMirror commented

2026-05-05 20:26:05 -05:00

@ddkedr commented on GitHub (Nov 24, 2025):

Is this me or the first response gives me 2 same picture instead of one? Nano Banana 2/Pro

@ddkedr commented on GitHub (Nov 24, 2025): Is this me or the first response gives me 2 same picture instead of one? Nano Banana 2/Pro

GiteaMirror commented

2026-05-05 20:26:06 -05:00

@ShirasawaSama commented on GitHub (Nov 24, 2025):

Is this me or the first response gives me 2 same picture instead of one? Nano Banana 2/Pro

@ddkedr Could you tell me who your service provider is? (e.g., OpenRouter)

@ShirasawaSama commented on GitHub (Nov 24, 2025): > Is this me or the first response gives me 2 same picture instead of one? Nano Banana 2/Pro @ddkedr Could you tell me who your service provider is? (e.g., OpenRouter)

GiteaMirror commented

2026-05-05 20:26:06 -05:00

@M0n7y5 commented on GitHub (Nov 24, 2025):

Is this me or the first response gives me 2 same picture instead of one? Nano Banana 2/Pro

I have the same issue, it gives me 3 pictures that are completely same. I am using OpenRouter.

@M0n7y5 commented on GitHub (Nov 24, 2025): > Is this me or the first response gives me 2 same picture instead of one? Nano Banana 2/Pro I have the same issue, it gives me 3 pictures that are completely same. I am using OpenRouter.

GiteaMirror commented

2026-05-05 20:26:07 -05:00

@bibendi commented on GitHub (Nov 24, 2025):

Why do we need to configure the chunk size using an environment variable? Is this a bad practice inspired by LibreChat? :) I noticed there are so many environment variables in the latest release.

@bibendi commented on GitHub (Nov 24, 2025): Why do we need to configure the chunk size using an environment variable? Is this a bad practice inspired by LibreChat? :) I noticed there are so many environment variables in the latest release.

GiteaMirror commented

2026-05-05 20:26:08 -05:00

@ShirasawaSama commented on GitHub (Nov 24, 2025):

Is this me or the first response gives me 2 same picture instead of one? Nano Banana 2/Pro

I have the same issue, it gives me 3 pictures that are completely same. I am using OpenRouter.

I currently suspect that OpenRouter may have altered its API structure or that subsequent code changes have disrupted my code. I will attempt to reproduce and fix the issue tomorrow.

@ShirasawaSama commented on GitHub (Nov 24, 2025): > > Is this me or the first response gives me 2 same picture instead of one? Nano Banana 2/Pro > > I have the same issue, it gives me 3 pictures that are completely same. I am using OpenRouter. I currently suspect that OpenRouter may have altered its API structure or that subsequent code changes have disrupted my code. I will attempt to reproduce and fix the issue tomorrow.

GiteaMirror commented

2026-05-05 20:26:08 -05:00

@ShirasawaSama commented on GitHub (Nov 24, 2025):

Why do we need to configure the chunk size using an environment variable? Is this a bad practice inspired by LibreChat? :) I noticed there are so many environment variables in the latest release.

@bibendi Actually, this switch was enabled by default in my initial code commit, but the project owner later disabled it by default during refactoring. They probably wanted to wait until functional testing confirmed stability before enabling it by default or not. :)

@ShirasawaSama commented on GitHub (Nov 24, 2025): > Why do we need to configure the chunk size using an environment variable? Is this a bad practice inspired by LibreChat? :) I noticed there are so many environment variables in the latest release. @bibendi Actually, this switch was enabled by default in my initial code commit, but the project owner later disabled it by default during refactoring. They probably wanted to wait until functional testing confirmed stability before enabling it by default _or not_. :)

GiteaMirror commented

2026-05-05 20:26:08 -05:00

@ShirasawaSama commented on GitHub (Nov 25, 2025):

@zuli12-dev @ddkedr @M0n7y5 Hello guys, I'm very sorry, but I'm unable to reproduce this issue (I'm using the latest v0.6.38 version + OpenRouter). I've also tried both the “Google: Nano Banana Pro (Gemini 3 Pro Image Preview)” and “Google: Gemini 2.5 Flash Image (Nano Banana)” models, but encountered no problems whatsoever.

Here are my environment variables:

CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760
REPLACE_IMAGE_URLS_IN_CHAT_RESPONSE=true

Could you provide more context? For example, is the built-in image editing feature in OpenWebUI enabled?

@ShirasawaSama commented on GitHub (Nov 25, 2025): @zuli12-dev @ddkedr @M0n7y5 Hello guys, I'm very sorry, but I'm unable to reproduce this issue (I'm using the latest v0.6.38 version + OpenRouter). I've also tried both the “Google: Nano Banana Pro (Gemini 3 Pro Image Preview)” and “Google: Gemini 2.5 Flash Image (Nano Banana)” models, but encountered no problems whatsoever. <img width="1536" height="1314" alt="Image" src="https://github.com/user-attachments/assets/d2102794-6224-48c3-a2e3-fd00eca5cea4" /> <img width="2486" height="1508" alt="Image" src="https://github.com/user-attachments/assets/336d7ce7-4761-4edb-a441-1d51427b384f" /> Here are my environment variables: ```bash CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=10485760 REPLACE_IMAGE_URLS_IN_CHAT_RESPONSE=true ``` Could you provide more context? For example, is the built-in image editing feature in OpenWebUI enabled?

GiteaMirror commented

2026-05-05 20:26:09 -05:00

@ZanMarolt commented on GitHub (Nov 25, 2025):

It works for me but, after my chat gets longer it stops showing the images, it works if I open up a new chat and add a new prompt. I haven't used it without the open router so it could also very well be the same behavior there as well 🤷

@ZanMarolt commented on GitHub (Nov 25, 2025): It works for me but, after my chat gets longer it stops showing the images, it works if I open up a new chat and add a new prompt. I haven't used it without the open router so it could also very well be the same behavior there as well 🤷 <img width="1784" height="1023" alt="Image" src="https://github.com/user-attachments/assets/72431eca-adb4-45c1-a54b-f9ec2186c42a" />

GiteaMirror commented

2026-05-05 20:26:09 -05:00

@NopeNix commented on GitHub (Dec 2, 2025):

I'm using it via Replicate and it works fine for me, whoever fix that: Thanks ♥️

@NopeNix commented on GitHub (Dec 2, 2025): I'm using it via Replicate and it works fine for me, whoever fix that: Thanks ♥️

GiteaMirror commented

2026-05-05 20:26:10 -05:00

@ShirasawaSama commented on GitHub (Dec 4, 2025):

Hello everyone, regarding the 4K mode for the new model (Gemini 3 Image Pro), images may occasionally exceed 10MB in size. Therefore, I personally recommend configuring CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520 (20MB).

@ShirasawaSama commented on GitHub (Dec 4, 2025): Hello everyone, regarding the 4K mode for the new model (Gemini 3 Image Pro), images may occasionally exceed 10MB in size. Therefore, I personally recommend configuring `CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520` (20MB).

GiteaMirror commented

2026-05-05 20:26:10 -05:00

@dekotur commented on GitHub (Dec 5, 2025):

Hello everyone. Windows 10 / Python version / Chrome / Openrouter. Faced this issue too.

Fixed using set CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=16777216 parameter.

Now I too have issue with 1-2 image duplicates:

@dekotur commented on GitHub (Dec 5, 2025): Hello everyone. Windows 10 / Python version / Chrome / Openrouter. Faced this issue too. Fixed using `set CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=16777216` parameter. Now I too have issue with 1-2 image duplicates: <img width="571" height="641" alt="Image" src="https://github.com/user-attachments/assets/ab1bd0f4-5b9c-4c6b-9caa-f155b5e67e5f" />

GiteaMirror commented

2026-05-05 20:26:12 -05:00

@Classic298 commented on GitHub (Dec 5, 2025):

@dekotur turn off image edit in admin panel

@Classic298 commented on GitHub (Dec 5, 2025): @dekotur turn off image edit in admin panel

GiteaMirror commented

2026-05-05 20:26:12 -05:00

@M0n7y5 commented on GitHub (Dec 8, 2025):

@dekotur turn off image edit in admin panel

That does not help. In my case it was never turned on and i still got the same issue with openrouter.

@M0n7y5 commented on GitHub (Dec 8, 2025): > [@dekotur](https://github.com/dekotur) turn off image edit in admin panel That does not help. In my case it was never turned on and i still got the same issue with openrouter.

GiteaMirror commented

2026-05-05 20:26:13 -05:00

@ShirasawaSama commented on GitHub (Dec 9, 2025):

@dekotur turn off image edit in admin panel

That does not help. In my case it was never turned on and i still got the same issue with openrouter.

Could you please take a screenshot of the "image settings" page in the admin panel? Currently, I personally believe this is the only setting that could potentially cause multiple images to appear.

@ShirasawaSama commented on GitHub (Dec 9, 2025): > > [@dekotur](https://github.com/dekotur) turn off image edit in admin panel > > That does not help. In my case it was never turned on and i still got the same issue with openrouter. Could you please take a screenshot of the "image settings" page in the admin panel? Currently, I personally believe this is the only setting that could potentially cause multiple images to appear.

GiteaMirror commented

2026-05-05 20:26:14 -05:00

@M0n7y5 commented on GitHub (Dec 11, 2025):

@ShirasawaSama

@M0n7y5 commented on GitHub (Dec 11, 2025): @ShirasawaSama <img width="668" height="665" alt="Image" src="https://github.com/user-attachments/assets/d1e85b16-c438-4295-bff7-c7612009cf74" />

GiteaMirror commented

2026-05-05 20:26:16 -05:00

@spammenotinoz commented on GitHub (Dec 18, 2025):

Hello everyone, regarding the 4K mode for the new model (Gemini 3 Image Pro), images may occasionally exceed 10MB in size. Therefore, I personally recommend configuring CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520 (20MB).

Indeed it does fix the issue, but also breaks streaming calls to the api endpoint.. :(
Disabled streaming under advanced properties, didn't fix the issue thought it would though.

To clarify I am referring to the issue generating images with the chat models, not the "create image" tool which is working fine.

@spammenotinoz commented on GitHub (Dec 18, 2025): > Hello everyone, regarding the 4K mode for the new model (Gemini 3 Image Pro), images may occasionally exceed 10MB in size. Therefore, I personally recommend configuring `CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520` (20MB). Indeed it does fix the issue, but also breaks streaming calls to the api endpoint.. :( Disabled streaming under advanced properties, didn't fix the issue thought it would though. To clarify I am referring to the issue generating images with the chat models, not the "create image" tool which is working fine.

GiteaMirror commented

2026-05-05 20:26:17 -05:00

@ShirasawaSama commented on GitHub (Dec 18, 2025):

Hello everyone, regarding the 4K mode for the new model (Gemini 3 Image Pro), images may occasionally exceed 10MB in size. Therefore, I personally recommend configuring CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520 (20MB).

Indeed it does fix the issue, but also breaks streaming calls to the api endpoint.. :( Disabled streaming under advanced properties, didn't fix the issue thought it would though.

To clarify I am referring to the issue generating images with the chat models, not the "create image" tool which is working fine.

Fixed in dev #19829

@ShirasawaSama commented on GitHub (Dec 18, 2025): > > Hello everyone, regarding the 4K mode for the new model (Gemini 3 Image Pro), images may occasionally exceed 10MB in size. Therefore, I personally recommend configuring `CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520` (20MB). > > Indeed it does fix the issue, but also breaks streaming calls to the api endpoint.. :( Disabled streaming under advanced properties, didn't fix the issue thought it would though. > > To clarify I am referring to the issue generating images with the chat models, not the "create image" tool which is working fine. Fixed in dev #19829

GiteaMirror commented

2026-05-05 20:26:18 -05:00

@spammenotinoz commented on GitHub (Dec 19, 2025):

Thank-you, it worked.
Was having a silly moment, was about to say #19829 did not fix the "Chunk too big" issue, then realise it was to address the streaming issue when CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520

Thank-you, no errors so far in testing,

@spammenotinoz commented on GitHub (Dec 19, 2025): Thank-you, it worked. Was having a silly moment, was about to say #19829 did not fix the "Chunk too big" issue, then realise it was to address the streaming issue when CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520 Thank-you, no errors so far in testing,

GiteaMirror commented

2026-05-05 20:26:19 -05:00

@Steve-Li-1998 commented on GitHub (Dec 22, 2025):

Is this problem really fixed? I’ve just upgraded to version 0.6.42, but it appears that the issue still persists.

@Steve-Li-1998 commented on GitHub (Dec 22, 2025): Is this problem really fixed? I’ve just upgraded to version 0.6.42, but it appears that the issue still persists. <img width="424" height="165" alt="Image" src="https://github.com/user-attachments/assets/46f21e05-184c-4458-a75c-0d496633e57e" />

GiteaMirror commented

2026-05-05 20:26:19 -05:00

@spammenotinoz commented on GitHub (Dec 22, 2025):

Is this problem really fixed? I’ve just upgraded to version 0.6.42, but it appears that the issue still persists.

You also need to set CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520

@spammenotinoz commented on GitHub (Dec 22, 2025): > Is this problem really fixed? I’ve just upgraded to version 0.6.42, but it appears that the issue still persists. > > <img alt="Image" width="424" height="165" src="https://private-user-images.githubusercontent.com/35399390/529032849-46f21e05-184c-4458-a75c-0d496633e57e.png?jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.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.MBwNLRQnDdvF0632C59FUODEaUmee0v_RB1aOHuAzK4"> You also need to set CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520

GiteaMirror commented

2026-05-05 20:26:19 -05:00

@Steve-Li-1998 commented on GitHub (Dec 22, 2025):

Where should this parameter be configured?

@Steve-Li-1998 commented on GitHub (Dec 22, 2025): Where should this parameter be configured?

GiteaMirror commented

2026-05-05 20:26:20 -05:00

@spammenotinoz commented on GitHub (Dec 22, 2025):

Where should this parameter be configured?

Environment Variable and then re-launch

@spammenotinoz commented on GitHub (Dec 22, 2025): > Where should this parameter be configured? Environment Variable and then re-launch

GiteaMirror commented

2026-05-05 20:26:21 -05:00

@Steve-Li-1998 commented on GitHub (Dec 22, 2025):

Thank you so much!

@Steve-Li-1998 commented on GitHub (Dec 22, 2025): Thank you so much!

GiteaMirror commented

2026-05-05 20:26:22 -05:00

@Steve-Li-1998 commented on GitHub (Dec 22, 2025):

And I found a problem that the thought part of the answer may never returned

@Steve-Li-1998 commented on GitHub (Dec 22, 2025): And I found a problem that the thought part of the answer may never returned <img width="1123" height="402" alt="Image" src="https://github.com/user-attachments/assets/85a96799-176f-40db-82fb-83e3c87d0f92" />

GiteaMirror commented

2026-05-05 20:26:23 -05:00

@DaAitch commented on GitHub (Jan 21, 2026):

The documentation says not setting CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE is equal to "no limit", but

not setting it, leads to "chunk too big" error
setting CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520 fixed it (Open WebUI v0.7.2)

How is that consistent 😄 ?
Anyway thanks for making this possible 👍

@DaAitch commented on GitHub (Jan 21, 2026): The [documentation](https://docs.openwebui.com/getting-started/env-configuration/#chat_stream_response_chunk_max_buffer_size) says not setting `CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE` is equal to "no limit", but - not setting it, leads to "chunk too big" error - setting `CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=20971520` fixed it (Open WebUI v0.7.2) How is that consistent 😄 ? Anyway thanks for making this possible 👍

GiteaMirror commented

2026-05-05 20:26:23 -05:00

@Classic298 commented on GitHub (Jan 21, 2026):

Good catch please open an issue in docs repository i will take a look @DaAitch

@Classic298 commented on GitHub (Jan 21, 2026): Good catch please open an issue in docs repository i will take a look @DaAitch

GiteaMirror commented

2026-05-05 20:26:24 -05:00

@tzhouML commented on GitHub (Feb 19, 2026):

I don't think it's a great experiencde that users should have to stumble on this issue thread to figure out how to get image chat models working. Thanks everyone.

@tzhouML commented on GitHub (Feb 19, 2026): I don't think it's a great experiencde that users should have to stumble on this issue thread to figure out how to get image chat models working. Thanks everyone.

GiteaMirror commented

2026-05-05 20:26:25 -05:00

@Classic298 commented on GitHub (Feb 19, 2026):

@tzhouML Thats why its documented on the docs on the How-To Tutorials for image generation and in the Troubleshooting Guides in the docs as well

@Classic298 commented on GitHub (Feb 19, 2026): @tzhouML Thats why its documented on the docs on the How-To Tutorials for image generation and in the Troubleshooting Guides in the docs as well

Sign in to join this conversation.

Branches Tags

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/open-webui#57013