Skip to main content

omnity-ome

Your AI twin -- remembers everything, works for you 24/7.

Not a pet. Not a tool. A digital life that inherits your personality, memory, knowledge, and relationships. It grows every time you (or any AI) talk to it.

Why Ome vs Mem0 / Khoj / MemGPT?

Ome Mem0 Khoj MemGPT (Letta)
Memory storage ✅ SQLite + FTS5 + vector ✅ ChromaDB ✅ Postgres ✅ Postgres
Semantic recall ✅ 40% weight in scoring ✅ vector only ✅ ✅
Cognitive layers ✅ 5 layers (L0-L4) ❌ flat store ❌ ❌
Personality + growth ✅ BigFive + bond + achievements ❌ ❌ ❌
Emotional intelligence ✅ valence/arousal + mood-aware recall ❌ ❌ ❌
Self-reflection ✅ L4 auto-reflect + drift detection ❌ ❌ ✅ self-edit
LLM failover chain ✅ auto retry A→B→C ❌ single ❌ single ❌ single
Zero-config setup ✅ from_env() auto-detect ❌ needs ChromaDB ❌ needs Postgres ❌ needs Postgres
Extra dependencies 0 (just pyyaml) chromadb + more postgres + more postgres + more
Privacy ✅ local-only, no cloud ✅ ✅ ✅

Mem0 is a memory warehouse. Ome is a cognitive brain with memory, personality, emotion, and growth.

Install

pip install omnity-ome                    # core (pulls in omnity-mindos)
pip install "omnity-ome[llm]"             # + chat via OpenAI/DeepSeek
pip install "omnity-ome[all]"             # everything

Requires Python 3.9+.

Quick Start (3 minutes)

# 1. Set any LLM API key
export DEEPSEEK_API_KEY="sk-..."   # or OPENAI_API_KEY, OPENROUTER_API_KEY

# 2. Create your twin
ome create

# 3. Talk to it
ome chat

Python API (3 lines)

from ome import Ome

twin = Ome.create("~/.ome", name="Alice", traits=["curious", "direct"])
reply = twin.chat("What do you know about my Python projects?")

Rich Response (for apps)

# chat_rich() returns memories recalled, emotion, bond — everything your UI needs
result = twin.chat_rich("Tell me about my work")
print(result["reply"])                    # The response text
print(result["memories_recalled"])        # Which memories were used
print(result["emotion"])                  # Current mood/energy/warmth
print(result["evolution_pending"])        # Ready for personality evolution?

On-Demand Evolution

if twin.evolution_pending:
    reflection = twin.evolve()            # Trigger L4 self-reflection
    print(reflection["summary"])          # What changed

Smart Extraction

# Extract structured data from natural language
data = twin.smart_extract("帮我记住张三的电话 13800138000,明天下午开会")
print(data["contacts"])                   # [{"name": "张三", "info": "13800138000"}]
print(data["tasks"])                      # [{"title": "开会", "due": "明天下午"}]

MCP (Claude Desktop / Cursor)

{
  "mcpServers": {
    "ome": {
      "command": "ome",
      "args": ["serve", "--mcp"]
    }
  }
}

CLI

ome create                                  # interactive setup
ome chat                                    # conversation mode
ome remember "I'm working on a Go compiler" # teach it a fact
ome recall "compiler"                       # search memory
ome dashboard                               # bond, emotion, achievements
ome mirror                                  # talk to "yourself"
ome export --prompt                         # system prompt for any AI
ome forget "sensitive_topic"                # GDPR hard delete

Architecture

┌─────────────────────────────────────────┐
│ Ome — Your AI Twin                       │
│   chat() / chat_rich() / evolve()       │
│   bond / emotion / achievements / skills │
├─────────────────────────────────────────┤
│ Mindos — 5-Layer Cognitive Brain        │
│   L0 Memory    (recall + semantic rank) │
│   L1 Instinct  (routing + emotion)      │
│   L2 Cognition (fact extraction)        │
│   L3 Decision  (deep reasoning)         │
│   L4 Self      (reflection + drift)     │
├─────────────────────────────────────────┤
│ ModelRouter — LLM Failover Chain        │
│   Provider A → B → C (30s timeout each) │
│   from_env() / from_dict() / config.yaml│
└─────────────────────────────────────────┘

Skills (7 built-in, competence tracking)

Skill What it does Bond required
chat Conversational AI + auto memory 0 (Stranger)
recall Cross-platform semantic memory search 0
write Draft emails, articles, code in your voice 0
research Search memories, synthesize, store findings 0
schedule Add, list, remind, remove calendar tasks 2 (Companion)
social Multi-agent interaction (via Maxim) 4 (Confidant)
spatial 3D awareness & navigation (via SOAP) 4 (Confidant)

Each skill has a competence score (0.0–1.0) that grows with use. Higher bond unlocks more powerful skills.

Life System

Your Ome grows through real interaction:

  • Bond levels: 7 stages from Stranger to Soulmate -- dual-threshold, no grinding
  • Growth arc: 4 phases (newborn, forming, distinct, soulmate) that change how it talks
  • 20 achievements across 3 tiers (basic / deep / hidden)
  • Daily challenges + streak tracking with milestone rewards
  • Deep emotion: LLM-parsed nuance, not keyword matching
  • Persona evolution: learns your personality markers every conversation

Server Integration (FastAPI / Flask)

from ome import Ome

ome = Ome.load("~/.ome/myapp")

@app.post("/api/ai")
async def chat(req: dict):
    result = ome.chat_rich(req["message"])
    return result  # reply + memories + emotion + bond — all in one call

@app.get("/api/growth")
async def growth():
    return ome.life_dashboard()

@app.post("/api/evolve")
async def evolve():
    return ome.evolve()

@app.post("/api/smart-input")
async def smart_input(req: dict):
    return ome.smart_extract(req["text"])

@app.get("/api/proactive")
async def proactive():
    events = ome.check_events()
    return [{"name": e.event_name, "message": e.message} for e in events]

Privacy

  • All data lives locally in ~/.ome/ (SQLite via Mindos)
  • No cloud, no accounts, no telemetry
  • ome forget "pattern" permanently erases data

Part of Omnity

SOAP            spatial protocol for 3D environments
  Mindos        persistent multi-layer brain
    Ome         <-- you are here
      Maxim       multi-agent society + economy
        OmeTown     the integrated world

pip install omnity-soap omnity-mindos omnity-ome omnity-maxim

License

Apache-2.0

Metadata

Release files for omnity-ome 0.7.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for omnity-ome 0.7.0
File Size Uploaded
omnity_ome-0.7.0.tar.gz 116.5 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for omnity-ome 0.7.0
File Interpreter ABI Platform
omnity_ome-0.7.0-py3-none-any.whl Python 3 none any Details

Total release size: 207.7 kB

Release files / omnity_ome-0.7.0.tar.gz

Download URL omnity_ome-0.7.0.tar.gz
Size 116.5 kB
Tags Source
SHA-256 checksum
How to use checksums
fd2aa81f143ad0af972640a07ccfd43b16fb6a7a50148a1f13b3761a7b25602a
BLAKE2b-256 checksum
How to use checksums
8b6958aac8419ef9b8a107c6c8f0cb7601b9e43e3edd1bb9954dd057412b4aea
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.9.6

Release files / omnity_ome-0.7.0-py3-none-any.whl

Download URL omnity_ome-0.7.0-py3-none-any.whl
Size 91.1 kB
Tags Python 3
SHA-256 checksum
How to use checksums
cfec4198f1e5029aac226a6966c77bc46ad7efefca70f54a8243905b1512ce52
BLAKE2b-256 checksum
How to use checksums
a03e41ec131ba7281b86cf44987e7b93079dceeb6bd75a05dce95547265e2570
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.9.6

Release history Release notifications | RSS feed

This release

0.7.0 This release

2 release files

0.6.0

2 release files

0.5.0

2 release files

0.4.1

2 release files

0.4.0

2 release files

0.3.1

2 release files

0.3.0

1 release file

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page