github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #3824] Server error when submitting a request through OpenAI client #2367

New Issue
Closed
opened 2026-04-12 12:41:05 -05:00 by GiteaMirror · 7 comments
GiteaMirror commented 2026-04-12 12:41:05 -05:00
Owner
Copy Link

Originally created by @mishushakov on GitHub (Apr 22, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/3824

What is the issue?

Using Vercel AI SDK and llama2

Request:

const content = [
    { type: 'text', text: prompt },
    { type: 'text', text: page.content },
 ]

const result = await experimental_generateObject({
    model,
    schema,
    messages: [{ role: 'user', content }],
    temperature,
  })

Response:

mish@Mishs-MBP llm-scraper % bun examples/hn.ts                                                            
12 |     statusCode === 429 || // too many requests
13 |     statusCode >= 500),
14 |     // server error
15 |     data
16 |   }) {
17 |     super(message);
         ^
AI_APICallError: json: cannot unmarshal array into Go struct field Message.messages.content of type string
 url: "http://127.0.0.1:11434/v1/chat/completions"

      at new APICallError (/Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider/dist/index.mjs:17:5)
      at /Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider-utils/dist/index.mjs:279:12
12 |     statusCode === 429 || // too many requests
13 |     statusCode >= 500),
14 |     // server error
15 |     data
16 |   }) {
17 |     super(message);
         ^
AI_APICallError: json: cannot unmarshal array into Go struct field Message.messages.content of type string
 url: "http://127.0.0.1:11434/v1/chat/completions"

      at new APICallError (/Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider/dist/index.mjs:17:5)
      at /Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider-utils/dist/index.mjs:279:12

Possibly related to #3690?

OS

macOS

GPU

Apple

CPU

Apple

Ollama version

0.1.32

Originally created by @mishushakov on GitHub (Apr 22, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/3824 ### What is the issue? Using Vercel AI SDK and llama2 Request: ```ts const content = [ { type: 'text', text: prompt }, { type: 'text', text: page.content }, ] const result = await experimental_generateObject({ model, schema, messages: [{ role: 'user', content }], temperature, }) ``` Response: ``` mish@Mishs-MBP llm-scraper % bun examples/hn.ts 12 | statusCode === 429 || // too many requests 13 | statusCode >= 500), 14 | // server error 15 | data 16 | }) { 17 | super(message); ^ AI_APICallError: json: cannot unmarshal array into Go struct field Message.messages.content of type string url: "http://127.0.0.1:11434/v1/chat/completions" at new APICallError (/Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider/dist/index.mjs:17:5) at /Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider-utils/dist/index.mjs:279:12 12 | statusCode === 429 || // too many requests 13 | statusCode >= 500), 14 | // server error 15 | data 16 | }) { 17 | super(message); ^ AI_APICallError: json: cannot unmarshal array into Go struct field Message.messages.content of type string url: "http://127.0.0.1:11434/v1/chat/completions" at new APICallError (/Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider/dist/index.mjs:17:5) at /Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider-utils/dist/index.mjs:279:12 ``` Possibly related to #3690? ### OS macOS ### GPU Apple ### CPU Apple ### Ollama version 0.1.32
GiteaMirror added the bug label 2026-04-12 12:41:05 -05:00
GiteaMirror commented 2026-04-12 12:41:06 -05:00
Author
Owner
Copy Link

@thinkverse commented on GitHub (Apr 22, 2024):

Ollama's OpenAI compatibility doesn't support content arrays, only text contents, you can read about the supported request fields in the documentation: https://github.com/ollama/ollama/blob/main/docs/openai.md#supported-request-fields.

If you want to send more messages, you must send them separately.

[
    {
        "role": "system",
        "content": "You are a helpful assistant."
    },
    {
        "role": "user",
        "content": "Hello!"
    }
]
<!-- gh-comment-id:2069746422 --> @thinkverse commented on GitHub (Apr 22, 2024): Ollama's OpenAI compatibility doesn't support content arrays, only text contents, you can read about the supported request fields in the documentation: https://github.com/ollama/ollama/blob/main/docs/openai.md#supported-request-fields. If you want to send more messages, you must send them separately. ```json [ { "role": "system", "content": "You are a helpful assistant." }, { "role": "user", "content": "Hello!" } ] ```
GiteaMirror commented 2026-04-12 12:41:06 -05:00
Author
Owner
Copy Link

@mishushakov commented on GitHub (Apr 22, 2024):

Alright, will keep that in mind 😉

<!-- gh-comment-id:2070176682 --> @mishushakov commented on GitHub (Apr 22, 2024): Alright, will keep that in mind 😉
GiteaMirror commented 2026-04-12 12:41:07 -05:00
Author
Owner
Copy Link

@mishushakov commented on GitHub (Apr 22, 2024):

I'm getting this now:

mish@Mishs-MBP llm-scraper % bun examples/hn.ts
12 |     statusCode === 429 || // too many requests
13 |     statusCode >= 500),
14 |     // server error
15 |     data
16 |   }) {
17 |     super(message);
         ^
AI_APICallError: json: cannot unmarshal array into Go struct field Message.messages.content of type string
 url: "http://127.0.0.1:11434/v1/chat/completions"

      at new APICallError (/Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider/dist/index.mjs:17:5)
      at /Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider-utils/dist/index.mjs:279:12
12 |     statusCode === 429 || // too many requests
13 |     statusCode >= 500),
14 |     // server error
15 |     data
16 |   }) {
17 |     super(message);
         ^
AI_APICallError: json: cannot unmarshal array into Go struct field Message.messages.content of type string
 url: "http://127.0.0.1:11434/v1/chat/completions"

      at new APICallError (/Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider/dist/index.mjs:17:5)
      at /Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider-utils/dist/index.mjs:279:12

Code:

const result = await experimental_generateObject({
    model,
    schema,
    messages: [{ role: 'assistant', content: '' }, { role: 'user', content: '' }],
    temperature,
 })
<!-- gh-comment-id:2070868418 --> @mishushakov commented on GitHub (Apr 22, 2024): I'm getting this now: ``` mish@Mishs-MBP llm-scraper % bun examples/hn.ts 12 | statusCode === 429 || // too many requests 13 | statusCode >= 500), 14 | // server error 15 | data 16 | }) { 17 | super(message); ^ AI_APICallError: json: cannot unmarshal array into Go struct field Message.messages.content of type string url: "http://127.0.0.1:11434/v1/chat/completions" at new APICallError (/Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider/dist/index.mjs:17:5) at /Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider-utils/dist/index.mjs:279:12 12 | statusCode === 429 || // too many requests 13 | statusCode >= 500), 14 | // server error 15 | data 16 | }) { 17 | super(message); ^ AI_APICallError: json: cannot unmarshal array into Go struct field Message.messages.content of type string url: "http://127.0.0.1:11434/v1/chat/completions" at new APICallError (/Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider/dist/index.mjs:17:5) at /Users/mish/Documents/Projects/llm-scraper/node_modules/@ai-sdk/provider-utils/dist/index.mjs:279:12 ``` Code: ```ts const result = await experimental_generateObject({ model, schema, messages: [{ role: 'assistant', content: '' }, { role: 'user', content: '' }], temperature, }) ```
GiteaMirror commented 2026-04-12 12:41:07 -05:00
Author
Owner
Copy Link

@thinkverse commented on GitHub (Apr 22, 2024):

From what I understand of Vercel AI SDK, it seems they convert messages.content into an array behind the scenes in the convertToLanguageModelPrompt function.

For instance, the if-statement on line 34 checks if the messages.content is a string, and if it is it'll convert it to an array.

Given that Ollama's OpenAI compatibility doesn't support that yet, Vercel's AI Core sadly isn't compatible either.

<!-- gh-comment-id:2071006813 --> @thinkverse commented on GitHub (Apr 22, 2024): From what I understand of Vercel AI SDK, it seems they convert `messages.content` into an array behind the scenes in the [convertToLanguageModelPrompt](https://github.com/vercel/ai/blob/c056579a34a7f71ebce87f0f83c3fc3da20e0a8e/packages/core/core/prompt/convert-to-language-model-prompt.ts#L11) function. For instance, [the if-statement on line 34](https://github.com/vercel/ai/blob/c056579a34a7f71ebce87f0f83c3fc3da20e0a8e/packages/core/core/prompt/convert-to-language-model-prompt.ts#L34-L39) checks if the `messages.content` is a string, and if it is it'll convert it to an array. Given that Ollama's OpenAI compatibility doesn't support that yet, Vercel's AI Core sadly isn't compatible either.
GiteaMirror commented 2026-04-12 12:41:07 -05:00
Author
Owner
Copy Link

@mishushakov commented on GitHub (Apr 22, 2024):

I have started an issue there too. Since this is not an issue with Ollama per se., I'm closing.
Thank you for all your help 😄

<!-- gh-comment-id:2071021031 --> @mishushakov commented on GitHub (Apr 22, 2024): I have started an issue there too. Since this is not an issue with Ollama per se., I'm closing. Thank you for all your help 😄
GiteaMirror commented 2026-04-12 12:41:08 -05:00
Author
Owner
Copy Link

@MaxLeiter commented on GitHub (May 5, 2024):

I believe https://github.com/ollama/ollama/pull/2506 would fix this, along with improving overall OpenAI compatibility.

<!-- gh-comment-id:2094909731 --> @MaxLeiter commented on GitHub (May 5, 2024): I believe https://github.com/ollama/ollama/pull/2506 would fix this, along with improving overall OpenAI compatibility.
GiteaMirror commented 2026-04-12 12:41:08 -05:00
Author
Owner
Copy Link

@mishushakov commented on GitHub (May 6, 2024):

Glad to hear that!

<!-- gh-comment-id:2095578951 --> @mishushakov commented on GitHub (May 6, 2024): Glad to hear that!
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#2367
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?