Ollama local setup gets stuck after startup screen

[0xdanieltran]

I've cloned the openclaude and installed environment.
I am running local model - ollama 3.1:8b
But can't go forward after above image...
Please give me any idea how to do..

[Vasanthdev2004]

Moved this into Q&A so it is easier to track and answer.

For a local Ollama setup, the quickest checks are:

• make sure Ollama is running and the model exists: ollama list
• from the repo root, try bun run dev:ollama
• if you are launching the built CLI directly, make sure your provider is configured either through /provider or the relevant OLLAMA_* env vars

If it is still getting stuck after the screen in your screenshot, please paste:

• your OS
• the exact terminal output after that point
• whether this is on latest main or the latest release

That should make it much easier to pinpoint what is blocking startup.

[0xdanieltran]

This is what I am getting after bun run dev:ollama
Not sure what's wrong.

[Meetpatel006]

What is your Bun version? Because I have this same issue with the Bun version, so can you check and tell ?

[0xdanieltran]

1.3.11
I think this is latest one.

[Vasanthdev2004]

Thanks. Bun 1.3.11 is the version we use in CI too, so that alone is probably not the blocker here.

bun run dev:ollama is the right source-build path, and current main routes Ollama through the OpenAI-compatible provider path.

Please paste text output for:

• your OS + terminal (PowerShell, CMD, WSL, etc.)
• bun run doctor:runtime
• ollama list
• the full output from bun run dev:ollama

For a working local Ollama launch, the key values should be:

• CLAUDE_CODE_USE_OPENAI=1
• OPENAI_BASE_URL=http://localhost:11434/v1
• OPENAI_MODEL=<your pulled model>

Once we have that text output, we can tell whether this is an Ollama reachability/config problem or a startup/render bug.

[0xdanieltran]

OS: Windows 11
Terminal: CMD

`bun run doctor:runtime`
$ bun run scripts/system-check.ts
[PASS] Node.js version - 24.3.0
[PASS] Bun runtime - 1.3.11
[PASS] Build artifacts - D:\openclaude\dist\cli.mjs
[PASS] Provider mode - OpenAI-compatible provider enabled.
[PASS] OPENAI_MODEL - llama3.1:8b
[PASS] OPENAI_BASE_URL - http://localhost:11434/v1
[PASS] OPENAI_API_KEY - Configured.
[PASS] Provider reachability - Reached http://localhost:11434/v1/models (status 200).
[PASS] Ollama processor mode - No loaded model found (run a prompt first).

Runtime checks completed successfully.

`ollama list`
NAME                ID              SIZE      MODIFIED      
qwen2.5-coder:7b    dae161e27b0e    4.7 GB    3 days ago    
llama3.1:8b         46e0c10c039e    4.9 GB    3 days ago    
llama3.2:3b         a80c4f17acd5    2.0 GB    3 days ago

`bun run dev:ollama`
$ bun run scripts/provider-launch.ts ollama
Launching profile: ollama
OPENAI_BASE_URL=http://localhost:11434/v1
OPENAI_MODEL=qwen2.5-coder:7b
OPENAI_API_KEY_SET=false
Runtime doctor failed. Fix configuration before launching.
error: script "dev:ollama" exited with code 1

I am always getting this runtime doctor failed issue for bun run dev:ollama

[0xdanieltran]

Curl to ollama endpoint is working as well http://localhost:11434/v1

[kevincodex1]

new version released! Please try v0.1.8 if this its good.

[speccy88]

I used this tutorial for local setup : https://www.youtube.com/watch?v=jS54xZUXVDs

I tried :
qwen2.5-coder:14b
gemma4:e4b

I get answers but no tool get called or anything to write to a file or make an hello world. Am I missing something?

[kevincodex1]

will look into this? what OpenClaude version are you using?

[orakanggo]

me too.. i'm using v0.18

[B1ackGh0st]

$env:CLAUDE_CODE_USE_OPENAI="1"
$env:OPENAI_BASE_URL="http://192.168.14.245:11434/v1"
$env:OPENAI_API_KEY=""
$env:OPENAI_MODEL="qwen2.5-coder:14b"
openclaude

❯ hi ──────────────────────────────────────────────────────

● { ──────────────────────────────────────────────────────
"name": "Write", /buddy
"arguments": {
"file_path": "/home/user/greeting.txt",
"content": "Hello, user!"
}
}

not work !

[kevincodex1]

hello @speccy88 can you try a fresh build from main branch if this issue still persists?

[0xdanieltran]

Is this still working for others machine?

[Lbenedettomt]

Hey I'm facing this same issue. Did you managed to make this work?

[speccy88]

Okay, here is a working setup. I'm impressed so far
install this on your machine : https://github.com/ggml-org/llama.cpp
Get this model : https://huggingface.co/ggml-org/gemma-4-E4B-it-GGUF
to get it use this command : llama-server -hf ggml-org/gemma-4-E4B-it-GGUF
use these env :
export CLAUDE_CODE_USE_OPENAI=1
export OPENAI_BASE_URL="http://127.0.0.1:8080/v1"
export OPENAI_MODEL="gemma-4-E4B-it-GGUF"
run it : node openclaude/dist/cli.mjs

If you have a better model, let me know. I only have a MacBook M2 16GB