github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #10164] Tool call - Ollama enforces usage of string in enums for JSON Schema #32430

New Issue
Closed
opened 2026-04-22 13:42:19 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-22 13:42:19 -05:00
Owner
Copy Link

Originally created by @AdamStrojek on GitHub (Apr 7, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10164

Originally assigned to: @ParthSareen on GitHub.

What is the issue?

I am using this MCP Server for Todoist: https://github.com/abhiz123/todoist-mcp-server.

The todoist_create_task tool provides the following JSON schema:

{
	"properties": {
	  "content": {
		"description": "The content/title of the task",
		"type": "string"
	  },
	  "description": {
		"description": "Detailed description of the task (optional)",
		"type": "string"
	  },
	  "due_string": {
		"description": "Natural language due date like 'tomorrow', 'next Monday', 'Jan 23' (optional)",
		"type": "string"
	  },
	  "priority": {
		"description": "Task priority from 1 (normal) to 4 (urgent) (optional)",
		"enum": [1, 2, 3, 4],
		"type": "number"
	  }
	},
	"required": ["content"],
	"type": "object"
}

When I submit this schema to Ollama, I receive a 400 error with the following message: {"error":{"message":"json: cannot unmarshal number into Go struct field .tools.function.parameters.properties.enum of type string","type":"invalid_request_error","param":null,"code":null}}.

This MCP Server functions correctly with both the OpenAI and Gemini APIs, and the schema is accepted without issue.

I have verified that the schema adheres to the JSON Schema documentation, which states that enums can utilize any type and mix values: https://json-schema.org/underststanding-json-schema/reference/enum.

I have tested this with multiple models, and the issue is model-independent.

What needs to be done?

To resolve this, Ollama needs to relax its schema validation to accept any type as a key for enum types.

Relevant log output

ggml_metal_init: skipping kernel_cpy_f32_bf16                      (not supported)
ggml_metal_init: skipping kernel_cpy_bf16_f32                      (not supported)
ggml_metal_init: skipping kernel_cpy_bf16_bf16                     (not supported)
llama_kv_cache_init: kv_size = 8192, offload = 1, type_k = 'f16', type_v = 'f16', n_layer = 32, can_shift = 1
llama_kv_cache_init:      Metal KV buffer size =  1024.00 MiB
llama_init_from_model: KV self size  = 1024.00 MiB, K (f16):  512.00 MiB, V (f16):  512.00 MiB
llama_init_from_model:        CPU  output buffer size =     3.10 MiB
llama_init_from_model:      Metal compute buffer size =   428.00 MiB
llama_init_from_model:        CPU compute buffer size =    22.01 MiB
llama_init_from_model: graph nodes  = 1286
llama_init_from_model: graph splits = 2
time=2025-04-07T17:51:28.083+02:00 level=INFO source=server.go:619 msg="llama runner started in 1.01 seconds"
[GIN] 2025/04/07 - 17:51:28 | 200 |  1.186805584s |       127.0.0.1 | POST     "/api/generate"
[GIN] 2025/04/07 - 17:51:53 | 400 |     651.666µs |       127.0.0.1 | POST     "/v1/chat/completions"
[GIN] 2025/04/07 - 17:51:56 | 400 |     655.916µs |       127.0.0.1 | POST     "/v1/chat/completions"
[GIN] 2025/04/07 - 17:55:02 | 400 |      631.75µs |       127.0.0.1 | POST     "/v1/chat/completions"
[GIN] 2025/04/07 - 17:55:05 | 400 |     464.167µs |       127.0.0.1 | POST     "/v1/chat/completions"


For each code this message is generated:

{"error":{"message":"json: cannot unmarshal number into Go struct field .tools.function.parameters.properties.enum of type string","type":"invalid_request_error","param":null,"code":null}}

OS

macOS

GPU

Apple

CPU

Apple

Ollama version

ollama version is 0.6.4

Originally created by @AdamStrojek on GitHub (Apr 7, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10164 Originally assigned to: @ParthSareen on GitHub. ### What is the issue? I am using this MCP Server for Todoist: https://github.com/abhiz123/todoist-mcp-server. The `todoist_create_task` tool provides the following JSON schema: ```json { "properties": { "content": { "description": "The content/title of the task", "type": "string" }, "description": { "description": "Detailed description of the task (optional)", "type": "string" }, "due_string": { "description": "Natural language due date like 'tomorrow', 'next Monday', 'Jan 23' (optional)", "type": "string" }, "priority": { "description": "Task priority from 1 (normal) to 4 (urgent) (optional)", "enum": [1, 2, 3, 4], "type": "number" } }, "required": ["content"], "type": "object" } ``` When I submit this schema to Ollama, I receive a 400 error with the following message: `{"error":{"message":"json: cannot unmarshal number into Go struct field .tools.function.parameters.properties.enum of type string","type":"invalid_request_error","param":null,"code":null}}`. This MCP Server functions correctly with both the OpenAI and Gemini APIs, and the schema is accepted without issue. I have verified that the schema adheres to the JSON Schema documentation, which states that enums can utilize any type and mix values: https://json-schema.org/underststanding-json-schema/reference/enum. I have tested this with multiple models, and the issue is model-independent. ## What needs to be done? To resolve this, Ollama needs to relax its schema validation to accept any type as a key for enum types. ### Relevant log output ```shell ggml_metal_init: skipping kernel_cpy_f32_bf16 (not supported) ggml_metal_init: skipping kernel_cpy_bf16_f32 (not supported) ggml_metal_init: skipping kernel_cpy_bf16_bf16 (not supported) llama_kv_cache_init: kv_size = 8192, offload = 1, type_k = 'f16', type_v = 'f16', n_layer = 32, can_shift = 1 llama_kv_cache_init: Metal KV buffer size = 1024.00 MiB llama_init_from_model: KV self size = 1024.00 MiB, K (f16): 512.00 MiB, V (f16): 512.00 MiB llama_init_from_model: CPU output buffer size = 3.10 MiB llama_init_from_model: Metal compute buffer size = 428.00 MiB llama_init_from_model: CPU compute buffer size = 22.01 MiB llama_init_from_model: graph nodes = 1286 llama_init_from_model: graph splits = 2 time=2025-04-07T17:51:28.083+02:00 level=INFO source=server.go:619 msg="llama runner started in 1.01 seconds" [GIN] 2025/04/07 - 17:51:28 | 200 | 1.186805584s | 127.0.0.1 | POST "/api/generate" [GIN] 2025/04/07 - 17:51:53 | 400 | 651.666µs | 127.0.0.1 | POST "/v1/chat/completions" [GIN] 2025/04/07 - 17:51:56 | 400 | 655.916µs | 127.0.0.1 | POST "/v1/chat/completions" [GIN] 2025/04/07 - 17:55:02 | 400 | 631.75µs | 127.0.0.1 | POST "/v1/chat/completions" [GIN] 2025/04/07 - 17:55:05 | 400 | 464.167µs | 127.0.0.1 | POST "/v1/chat/completions" For each code this message is generated: {"error":{"message":"json: cannot unmarshal number into Go struct field .tools.function.parameters.properties.enum of type string","type":"invalid_request_error","param":null,"code":null}} ``` ### OS macOS ### GPU Apple ### CPU Apple ### Ollama version ollama version is 0.6.4
GiteaMirror added the bug label 2026-04-22 13:42:20 -05:00
GiteaMirror commented 2026-04-22 13:42:22 -05:00
Author
Owner
Copy Link

@ParthSareen commented on GitHub (Apr 7, 2025):

Hey @AdamStrojek thanks for reporting! Will take a look!

<!-- gh-comment-id:2784077307 --> @ParthSareen commented on GitHub (Apr 7, 2025): Hey @AdamStrojek thanks for reporting! Will take a look!
GiteaMirror commented 2026-04-22 13:42:23 -05:00
Author
Owner
Copy Link

@ParthSareen commented on GitHub (Apr 8, 2025):

Will be fixed in next release :)

<!-- gh-comment-id:2787768470 --> @ParthSareen commented on GitHub (Apr 8, 2025): Will be fixed in next release :)
GiteaMirror commented 2026-04-22 13:42:25 -05:00
Author
Owner
Copy Link

@AdamStrojek commented on GitHub (Apr 9, 2025):

Great! Thank you!

<!-- gh-comment-id:2788380445 --> @AdamStrojek commented on GitHub (Apr 9, 2025): Great! Thank you!
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#32430
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?