Octopoda

The open-source memory operating system for AI agents.

Persistent memory, semantic search, loop detection, agent messaging, crash recovery, and real-time observability. Local-first. Works offline. Optionally sync to cloud.

Track latency, error rates, memory usage, and health scores per agent. Catch performance regressions before they hit production.

Browse every memory an agent has stored, search by prefix, and inspect the full version history of any key. See exactly how a memory changed over time — useful for debugging agent behavior and understanding why an agent made a decision.

Quick Start

pip install octopoda

from octopoda import AgentRuntime

agent = AgentRuntime("my_agent")
agent.remember("key", "value")
agent.recall("key")

That's it. Your agent now has persistent memory. It survives restarts, crashes, and deployments. Works locally with SQLite — no account required.

Want cloud sync and the dashboard? Just set an API key:

export OCTOPODA_API_KEY=sk-octopoda-...   # Get yours free at octopodas.com

Same code, now backed by PostgreSQL with real-time monitoring and multi-agent observability.

Why Octopoda

AI agents forget everything between sessions. Every framework treats memory as disposable. Octopoda gives your agents:

1. Persistent memory that survives restarts and crashes
2. Semantic search to find memories by meaning, not just exact keys
3. Loop detection that catches agents stuck in repetitive patterns
4. Agent messaging for multi-agent coordination
5. Audit trail so you can see every decision an agent made and why
6. Real-time dashboard to monitor what your agents know and how they're performing

Features

Everything works out of the box with pip install octopoda.

Semantic Search

agent.remember("bio", "Alice is a vegetarian living in London")
agent.remember("work", "Alice is a senior engineer at Google")

results = agent.recall_similar("where does the user work?")
# Returns: "Alice is a senior engineer at Google" (score: 0.82)

Loop Detection

Five signals: write similarity, key overwrites, velocity spikes, alert frequency, goal drift.

status = agent.get_loop_status()
# {"severity": "orange", "score": 45, "signals": [...]}
# Every signal tells you what's wrong and exactly what to do.

Agent Messaging

agent_a.send_message("agent_b", "Found a bug in auth", message_type="alert")
messages = agent_b.read_messages(unread_only=True)
agent_a.broadcast("Deploy starting in 5 minutes")

Goal Tracking

agent.set_goal("Migrate to PostgreSQL", milestones=["Backup", "Schema", "Migrate", "Validate"])
agent.update_progress(milestone_index=0, note="Backup done")
agent.get_goal()  # {"progress": 0.25, "milestones_completed": 1}

Memory Management

agent.forget("outdated_config")              # Delete specific memories
agent.forget_stale(days=30)                  # Remove old memories
agent.consolidate()                          # Merge duplicates
agent.memory_health()                        # {"score": 78, "issues": [...]}

Crash Recovery

agent.snapshot("before_migration")
# ... something goes wrong ...
agent.restore("before_migration")

Shared Memory

agent_a.share("research_pool", "analysis", {"findings": "..."})
data = agent_b.read_shared("research_pool", "analysis")

Export / Import

bundle = agent.export_memories()
new_agent.import_memories(bundle)

Framework Integrations

Drop-in memory for the frameworks you already use.

# LangChain
from synrix_runtime.integrations.langchain_memory import SynrixMemory
memory = SynrixMemory(agent_id="my_chain")

# CrewAI
from synrix_runtime.integrations.crewai_memory import SynrixCrewMemory
crew_memory = SynrixCrewMemory(crew_id="research_crew")

# AutoGen
from synrix_runtime.integrations.autogen_memory import SynrixAutoGenMemory
memory = SynrixAutoGenMemory(group_id="dev_team")

# OpenAI Agents SDK
from synrix.integrations.openai_agents import octopoda_tools
tools = octopoda_tools("my_agent")

MCP Server

Give Claude, Cursor, or any MCP-compatible AI persistent memory with zero code.

pip install octopoda[mcp]

Add to your Claude Desktop config (claude_desktop_config.json):

{
  "mcpServers": {
    "octopoda": {
      "command": "octopoda-mcp"
    }
  }
}

25 tools: memory operations, semantic search, loop detection, goal tracking, agent messaging, memory health, and more.

Cloud

Sign up at octopodas.com for the dashboard, managed hosting, and cloud API.

Setup:

export OCTOPODA_API_KEY=sk-octopoda-...

Or run octopoda-login to sign up interactively from your terminal.

from octopoda import Octopoda

client = Octopoda()  # Reads API key from env
agent = client.agent("my_agent")
agent.write("preference", "dark mode")
results = agent.search("user preferences")

Plans:

How It Compares

Installation

pip install octopoda              # Core (local memory, ~5 dependencies)
pip install octopoda[ai]          # + Local embeddings for semantic search
pip install octopoda[nlp]         # + spaCy for knowledge graph extraction
pip install octopoda[mcp]         # + MCP server for Claude/Cursor
pip install octopoda[all]         # Everything

Configuration

Contributing

See CONTRIBUTING.md for setup instructions and guidelines.

Security

See SECURITY.md for reporting vulnerabilities.

License

MIT — use it however you want. See LICENSE.

Built by RYJOX Technologies | PyPI | Cloud API | Dashboard