ZenCoderTroubleshooting
Common issues and how to fix them. If none of these help, check the daemon logs with ZENCODER_LOG_LEVEL=debug zencoderd.
Connection refused / daemon not reachable
Cause: zencoderd is not running, or is listening on a different address.
Start the daemon in a separate terminal:
zencoderd
# → ZenCoder daemon listening on :7777If you changed the address, set the env var in every terminal that runs the CLI or VSCode:
export ZENCODER_ADDR=http://myhost:8888
zencoder healthNo models appear in dropdown / zencoder models returns empty
Cause: Ollama isn't running, or no models have been pulled.
# Check Ollama is running
curl http://localhost:11434/api/tags
# Pull a model
ollama pull qwen2.5-coder:7b
# Verify the daemon can see it
zencoder modelsIf you're using LM Studio, make sure the Local Server is started in the app.
First response takes 30–60 seconds
Cause: The local model is being loaded into memory for the first time.
This is normal. Ollama keeps the model hot in memory after the first load. Subsequent requests will be much faster. Try a short warm-up:
zencoder chat "hi" # warms up the modelVS Code panel doesn't show after installing the extension
Cause: The extension may not have activated yet, or the command palette is the fastest way to open it.
Press Cmd+Shift+P (macOS) or Ctrl+Shift+P (Windows/Linux), type ZenCoder AI Chat, and press Enter.
If the command doesn't appear, try reloading VS Code with Cmd+Shift+P → Developer: Reload Window.
zencoder-secrets test returns 401 / unauthorized
Cause: The API key is wrong, expired, or doesn't have the required permissions.
Remove the key and re-add it, then test again:
# Remove the key
zencoder-secrets delete -provider anthropic
# Re-add it (you will be prompted to enter it securely)
zencoder-secrets add-byok --provider anthropic --name prod --model claude-3-haiku-4-5-20251001
# Test it
zencoder-secrets test -provider anthropicsk-ant-. OpenAI keys start with sk- or sk-proj-. NVIDIA NIM keys start with nvapi-.Context size warning / request rejected
Cause: The combined size of added files exceeds zenCoder.maxContextBytes.
Clear some context items, or raise the limit in VSCode settings:
// .vscode/settings.json
{
"zenCoder.maxContextBytes": 400000
}Alternatively, add individual files rather than whole folders.
Agent mode keeps iterating without finishing
Cause: The task is too broad, or the model is confused and repeating steps.
Hit Ctrl+C to stop the agent, then restart with a more specific task and a lower --max-iter limit:
> /mode agent --max-iter 5 "Add nil check to the GetUser handler in internal/server/handlers.go"Keep tasks scoped to a single file or function for best results.
permission denied reading ~/.zencoder/secrets/
Cause: The secrets directory has wrong permissions.
chmod 700 ~/.zencoder/secrets/
chmod 600 ~/.zencoder/secrets/*/model interactive picker shows 'requires a TTY'
Cause: You're running zencoder in a non-interactive shell (e.g. piped input, CI).
The interactive picker requires a terminal. Run zencoder directly in your terminal, then use:
> /model
# ↑↓ to navigate · Enter to select