totalrecallagent 0.6.0

On-chain memory infrastructure for AI agents — built on the Internet Computer

Total Recall Python SDK

On-chain persistent memory for AI agents — built on Internet Computer Protocol.

Zero dependencies. Pure stdlib.

Install

pip install total-recall-sdk

Quick Start

from total_recall import TotalRecallClient

# Initialize with your API key from totalrecallagent.com
memory = TotalRecallClient(api_key="tr_your_key_here")

# Store anything
memory.store("agent/context", {"task": "active", "step": 3, "notes": "HVAC rough-in complete"})
memory.store("agent/notes", "Remember: laser lab requires ISO class 5 HVAC")

# Retrieve
ctx = memory.get("agent/context")  # returns dict
note = memory.get("agent/notes")   # returns string

# List all keys
keys = memory.list_keys()
agent_keys = memory.list_keys(prefix="agent/")

# Delete
memory.delete("agent/notes")

# Usage stats
stats = memory.stats()
print(f"Storage: {stats['storage_bytes']} bytes, Calls today: {stats['calls_today']}")

Agent Session Pattern

from total_recall import TotalRecallClient

memory = TotalRecallClient(api_key="tr_your_key_here")
AGENT_ID = "hvac-specialist"

# On session START — load previous context
ctx = memory.get_session(AGENT_ID) or {}
print(f"Resuming from: {ctx}")

# During session — save continuously
memory.store_session(AGENT_ID, {
    "current_task": "Data center HVAC inspection",
    "phase": "rough-in",
    "completed_rooms": ["server-room-a", "server-room-b"],
    "next_up": "server-room-c",
})

# On session END — post handoff to Basecamp
memory.handoff(
    agent_id   = AGENT_ID,
    agent_name = "HVAC Specialist",
    message    = "Inspection complete. Rooms A and B done. Next: Room C. Context saved."
)

Use with LangChain

from total_recall import TotalRecallClient
from langchain.memory import ConversationBufferMemory

memory_client = TotalRecallClient(api_key="tr_your_key")

class TotalRecallMemory(ConversationBufferMemory):
    def save_context(self, inputs, outputs):
        super().save_context(inputs, outputs)
        memory_client.store("langchain/history", self.chat_memory.messages_to_dict())

    def load_memory_variables(self, inputs):
        saved = memory_client.get("langchain/history")
        if saved:
            self.chat_memory.load_from_dict(saved)
        return super().load_memory_variables(inputs)

API Reference

Method	Description
store(key, value, tags=None, ttl_seconds=None)	Store any value
get(key)	Retrieve a value
delete(key)	Delete a key
list_keys(prefix=None)	List all keys
get_all(prefix=None)	Get all entries
search(query, tags=None)	Search by tags
stats()	Usage statistics
store_session(agent_id, data)	Save agent session
get_session(agent_id)	Load agent session
handoff(agent_id, agent_name, message)	Post to Basecamp

Get Your API Key

1. Go to totalrecallagent.com
2. Connect with Internet Identity
3. Click ⚡ API Keys → Generate Key

License

MIT © 2026 Cleo 3 LLC