Healthcheck was failing because curl is not installed in the LiteLLM
container, causing Traefik to mark it as unhealthy and not route traffic.
Disabled healthcheck as Traefik doesn't require it for routing.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
LiteLLM was binding to localhost by default, making it unreachable
from Traefik reverse proxy. Added --host 0.0.0.0 parameter to allow
connections from the Docker network.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Added Traefik configuration to make LiteLLM accessible at llm.ai.pivoine.art
for use with @openai/codex CLI tool.
Changes:
- Added AI_LITELLM_TRAEFIK_HOST to arty.yml (llm.ai.pivoine.art)
- Updated ai/compose.yaml litellm service with full Traefik labels
- HTTP to HTTPS redirect
- SSL termination via Let's Encrypt
- Compression and security headers
This allows external tools like Codex to use Claude models via
OpenAI-compatible API endpoint.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Removed custom Dockerfile and SFTP function integration in favor of
the simpler REST API approach (webui-export.py).
Changes:
- Restored webui service to use official Open WebUI image
- Removed custom Dockerfile.webui (paramiko build)
- Removed ai/functions/save_to_disk.py SFTP function
- Removed SSH key and functions volume mounts
The REST API export script (webui-export.py) is a simpler and more
flexible solution that doesn't require Docker modifications.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Added custom Open WebUI function for SSH/SFTP file operations:
**New Function: save_to_disk.py**
- save_file(): Write generated code to local filesystem via SFTP
- read_file(): Read files from local disk
- list_files(): List directory contents
- Configurable via Valves (host, port, username, paths)
**Custom Dockerfile (Dockerfile.webui)**
- Based on ghcr.io/open-webui/open-webui:main
- Installs paramiko library for SSH/SFTP support
- Creates .ssh directory for key storage
**Configuration Updates**
- Mount SSH private key from host (/root/.ssh/id_rsa)
- Mount functions directory for custom tools
- Build custom image with SFTP capabilities
**Usage in Open WebUI**
Claude can now use these tools to:
- Generate code and save it directly to your local disk
- Read existing files for context
- List project directories
- Create new files in any project
Default base path: /home/valknar/Projects
Authentication: SSH key-based (passwordless)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Added LiteLLM as an OpenAI-compatible proxy for Anthropic's API to
enable Claude models in Open WebUI.
**New Service: litellm**
- Image: ghcr.io/berriai/litellm:main-latest
- Internal proxy on port 4000
- Converts Anthropic API to OpenAI-compatible format
- Health check with 30s intervals
- Not exposed via Traefik (internal only)
**LiteLLM Configuration (litellm-config.yaml)**
- Claude Sonnet 4 (claude-sonnet-4-20250514)
- Claude Sonnet 4.5 (claude-sonnet-4-5-20250929)
- Claude 3.5 Sonnet (claude-3-5-sonnet-20241022)
- Claude 3 Opus (claude-3-opus-20240229)
- Claude 3 Haiku (claude-3-haiku-20240307)
**Open WebUI Configuration Updates**
- Changed OPENAI_API_BASE_URLS to point to LiteLLM proxy
- URL: http://litellm:4000/v1
- Added litellm as dependency for webui service
- Dummy API key for proxy authentication
**Why LiteLLM?**
Anthropic's API uses different endpoint structure and authentication
headers compared to OpenAI. LiteLLM acts as a translation layer,
allowing Open WebUI to use Claude models through its OpenAI-compatible
interface.
**Available Models in Open WebUI**
- claude-sonnet-4 (latest Claude Sonnet 4)
- claude-sonnet-4.5 (Claude Sonnet 4.5)
- claude-3-5-sonnet
- claude-3-opus
- claude-3-haiku
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Changed service name from 'postgres' to 'ai_postgres' to avoid naming
conflict with the core PostgreSQL service in Docker Compose include.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Created complete AI infrastructure stack at ai.pivoine.art:
**New Services:**
- **Open WebUI** (ai.pivoine.art)
- ChatGPT-like interface for AI models
- Multi-user chat with authentication
- RAG (Retrieval-Augmented Generation) support
- Document upload and processing
- Claude API integration via Anthropic
- **PostgreSQL with pgvector** (dedicated AI database)
- Vector similarity search for RAG
- Separate from production databases
- Stores embeddings and documents
- **Crawl4AI** (internal API service)
- Web scraping optimized for LLMs
- Converts websites to clean Markdown
- Called by n8n workflows
- No public exposure (internal only)
**Configuration:**
- Added 18 AI environment variables to arty.yml
- Configured email notifications via IONOS SMTP
- OpenAI API compatibility for Claude integration
- Traefik SSL termination and compression
**Backup:**
- Added 3 AI volumes to Restic backup
- Daily backup at 3 AM
- Retention: 7 daily, 4 weekly, 6 monthly, 2 yearly
**Integration:**
- Shares falcon_network with existing services
- Ready for n8n workflow automation
- Mattermost notifications support
- Watchtower auto-updates enabled
Ready for Phase 2: GPU server integration with Ollama, Whisper, and
Stable Diffusion when IONOS A10 server is provisioned.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
