Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.sxth.ai/llms.txt

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

​
Mind

The Mind is the core abstraction in sxth-mind. It coordinates:
  • Adapter — Domain-specific behavior (identity types, journey stages, nudges)
  • Provider — LLM calls (OpenAI, custom, etc.)
  • Storage — Persistence (memory, SQLite, custom)

​
Creating a Mind

from sxth_mind import Mind
from sxth_mind.storage import SQLiteStorage
from sxth_mind.providers.openai import OpenAIProvider
from examples.sales import SalesAdapter

mind = Mind(
    adapter=SalesAdapter(),                     # Required
    provider=OpenAIProvider(),                  # Optional (defaults to OpenAI)
    storage=SQLiteStorage("mind.db"),           # Optional (defaults to memory)
)

​
Core Methods

​
chat()

The primary interface. Handles everything automatically:
  1. Loads or creates UserMind and ProjectMind
  2. Builds context-aware prompt using the adapter
  3. Calls the LLM
  4. Updates state based on the interaction
  5. Returns the response
response = await mind.chat(
    user_id="user_123",
    message="Following up with the enterprise lead",
    project_id="deal_acme",  # Optional, defaults to "default"
)

​
chat_stream()

Same as chat(), but yields tokens as they’re generated: 
async for token in mind.chat_stream("user_123", "Hello"):
    print(token, end="", flush=True)

​
get_state()

Retrieve the current cognitive state for introspection or custom use: 
state = await mind.get_state("user_123", project_id="deal_acme")

# Use state to personalize any LLM call
patterns = state["user_mind"]["patterns"]
journey_stage = state["project_mind"]["journey_stage"]

​
explain_state()

Get a human-readable summary of what the Mind knows: 
summary = await mind.explain_state("user_123")
print(summary)
# "User has logged 15 interactions. Pattern: enterprise deals need 4+ touches.
#  Current stage: negotiating (tone: strategic)"

​
get_pending_nudges()

Retrieve proactive suggestions for a user: 
nudges = await mind.get_pending_nudges("user_123")
for nudge in nudges:
    print(f"{nudge.title}: {nudge.message}")

​
Using State for Custom Personalization

You don’t have to use mind.chat(). Get state and use it however you want: 
# Get user understanding
state = await mind.get_state("user_123")
user_patterns = state["user_mind"]["patterns"]
journey_stage = state["project_mind"]["journey_stage"]

# Use it to personalize a dashboard
if journey_stage == "onboarding":
    show_tutorial_widgets()
elif user_patterns.get("prefers_detailed"):
    show_detailed_analytics()

# Use it in your own LLM calls
system_prompt = f"""
User is in the {journey_stage} stage.
Known patterns: {user_patterns}
Adapt your response accordingly.
"""
This makes sxth-mind useful for any AI-powered feature, not just chat.