Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.contextbase.dev/llms.txt

Use this file to discover all available pages before exploring further.

​
Basic usage

​
Initialize the client

from contextbase import Contextbase

# Using the CONTEXTBASE_API_KEY environment variable (recommended)
cb = Contextbase()

# Or pass API key directly
cb = Contextbase(api_key="your-api-key")

​
Publish JSON data

Publish structured data to build context memory: 
# Publish user interaction data
response = cb.publish(
    context_name="user_interactions",
    body={
        "user_message": "How do I reset my password?",
        "timestamp": "2024-01-15T14:30:00Z",
        "user_emotion": "confused",
        "session_id": "sess_123"
    },
    scopes={
        "user_id": 123,
        "project": "web-app"
    }
)

# Check if successful
if response.ok:
    print("✅ Data published successfully!")
else:
    print(f"❌ Error: {response.error.message}")

​
Publish files

Upload documents, images, and other files: 
from contextbase import ContextbaseFile

# Method 1: Simple file upload
response = cb.publish_file(
    context_name="user_uploads",
    file_path="user-manual.pdf",
    scopes={"project": "web-app"}
)

# Method 2: Using ContextbaseFile for more control
file = ContextbaseFile.from_path("image.png")
response = cb.publish(
    context_name="screenshots",
    file=file
)

# Method 3: Create file from string content
content = "Meeting notes: Discussed new features..."
file = ContextbaseFile.from_data(content, "notes.txt")
response = cb.publish(
    context_name="meeting_notes",
    file=file
)
See file handling

​
Resolve context

Get compressed, relevant context for your prompt: 
# Resolve with scopes and query
response = client.resolve_context(
    context_name="chatbot-assistant",
    scopes={"user_id": 123},
    query="password reset help"
)

if response.ok:
    context = response.json["content"]
    print(f"Context for AI: {context}")
    
    # Use this context in your AI prompt
    ai_prompt = f"""
    {context}
    
    User: How do I reset my password?
    Assistant: """

​
Using the decorator

Automatically publish function results: 
from contextbase import publish

# Publish function return values
@publish("analytics", "user-actions", scopes=lambda result: {"user_id": result["user_id"]})
def track_user_action(user_id, action):
    return {
        "user_id": user_id,
        "action": action,
        "timestamp": "2024-01-15T14:30:00Z"
    }

# Function runs normally and publishes result
track_user_action(123, "login")

# Publish as file
@publish("reports", "daily-summary", as_file=True, file_name="summary.txt")
def generate_daily_report():
    return "Daily Report: All systems operational"

# Creates and uploads summary.txt automatically
generate_daily_report()
More examples here

​
Error handling

Handle errors gracefully: 
from contextbase import ContextbaseError

try:
    response = cb.publish("context", "component", body={"test": "data"})
    response.raise_for_status()  # Raises exception if failed
    print("Success!")
except ContextbaseError as e:
    print(f"API Error: {e.message}")
    print(f"Status: {e.status_code}")
    for error in e.errors:
        print(f"- {error}")
except ValueError as e:
    print(f"Validation Error: {e}")

​
Next steps

Happy building! 🚀