This is an unofficial wrapper for Langdock API.

Notes:

• ✅ Fully implemented and working
• ⚠️ Implemented but has issues (HTTP.jl compatibility issue, works with curl)
• ❌ Not yet implemented

using Langdock

# Set up provider
ENV["LANGDOCK_API_KEY"] = "your-api-key"
provider = LangdockProvider(api_key=ENV["LANGDOCK_API_KEY"], region="eu") # or get_default_provider()

# List available assistant models
models = list_assistant_models(provider) # not working due to issue with HTTP.jl when calling this endpoint

# Create assistant chat
messages = [Dict("role" => "user", "content" => "Hello!")]
response = create_assistant_chat(
   provider,
   messages,
   assistant_id="<your-assistant-id>"
)

# Generate embeddings
embeddings = create_openai_embeddings(
   provider,
   "Text to embed",
   model_id="text-embedding-ada-002" 
)

To implement additional endpoints:

1. Create a new file in src/api/ for the endpoint category (if not exists)
2. Implement the function following the existing pattern:
   • Support both api_key and provider parameters
   • Use langdock_request for API calls
   • Add proper documentation with examples
3. Export the function in src/Langdock.jl
4. Add tests in the corresponding test/api/ file
5. Update this documentation

Priority endpoints to implement:

1. OpenAI Chat Completions - Most commonly used endpoint
2. Anthropic Messages - Support for Claude models
3. Knowledge Folder operations - File management capabilities
4. Mistral/Codestral - Code completion support

• Langdock API Documentation