Skip to main content

Orca — 100% Local Private AI

Your hardware. Your data. Your intelligence.
No Anthropic. No OpenAI. No cloud. No telemetry.

Orca is a fully private AI system that runs entirely on your own hardware using Ollama. It includes a terminal CLI, a professional web UI, a multi-agent Ultra mode, long-term memory, fine-tuning tools, and a self-contained revenue/licensing layer — all 100% local.


Quick Install

curl -fsSL https://orca.systems/install.sh | bash

Or via pip:

pip install orca-ai
orca doctor --wizard

Requirements

  • Python 3.11+
  • Ollama running locally
  • At least one Ollama model (e.g. ollama pull llama3.2:3b)

Getting Started

# First-run setup wizard
orca doctor --wizard

# Terminal chat
orca core chat

# Single-shot fast response
orca nano "explain recursion in 2 sentences"

# Web UI (opens in browser)
orca serve

# Multi-agent Ultra (Pro license required)
orca ultra run "design a REST API for a todo app"

Commands

Command Description
orca nano <prompt> Fast single-shot response
orca core chat Full interactive chat with memory + tools
orca core think <prompt> Deep single-shot reasoning
orca ultra run <task> Multi-agent orchestration
orca serve Launch the web UI
orca data seed --n 500 Generate synthetic training data
orca data curate Clean and score training data
orca train run Fine-tune via QLoRA
orca train cloud --ssh ... Train on a rented GPU
orca doctor System health check
orca doctor --wizard First-run setup wizard
orca upgrade Self-update from PyPI
orca activate <key> Activate a Pro license
orca license Show license status
orca status Live system dashboard

Features

Core

  • Full multi-turn chat with tool use (web search, code execution, file ops)
  • 4-layer memory: short-term, long-term (ChromaDB), episodic, semantic
  • Self-reflection and reasoning traces
  • Session save/resume

Ultra (Pro)

  • 6-agent parallel pipeline: researcher, coder, analyst, writer, critic, architect
  • Automatic decomposition, parallel execution, synthesis, grading, self-healing
  • Web UI pod visualization with live progress streaming

Fine-Tuning

  • Synthetic data generation across 20+ domains
  • QLoRA fine-tuning via Unsloth (local GPU)
  • Cloud training via SSH (Vast.ai, Lambda, RunPod)
  • GGUF export + Ollama registration

Web UI

  • Professional black-and-white design
  • CORE / ULTRA mode toggle
  • SSE streaming with real-time pod visualization
  • Memory recall sidebar
  • License status indicator

Licensing

Orca ships in two tiers:

Tier Price Features
Free $0 Core chat, doctor, status, data tools
Pro $49/mo + Ultra mode, cloud training, web UI
Enterprise $199/mo All features, 5 seats, priority support
orca activate ORCA-PRO-XXXXX-XXXXX-XXXXX
orca license --buy   # show pricing

Privacy

  • Zero telemetry
  • No external API calls from the core system
  • All data stored in ~/.orca/
  • Inference via Ollama on localhost:11434

Documentation

orca.systems/docs

Release files for orca-ai 0.1.1

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

Source distribution (sdist)

Source distribution for orca-ai 0.1.1
File Size Uploaded
orca_ai-0.1.1.tar.gz 102.4 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for orca-ai 0.1.1
File Interpreter ABI Platform
orca_ai-0.1.1-py3-none-any.whl Python 3 none any Details

Total release size: 234.3 kB

Release files / orca_ai-0.1.1.tar.gz

Download URL orca_ai-0.1.1.tar.gz
Size 102.4 kB
Tags Source
SHA-256 checksum
How to use checksums
d608b3d589f2ab645dd849963bb395e2717b86c59aa8020000cd0515ecdd4f4e
BLAKE2b-256 checksum
How to use checksums
1851c17de906089fdd1f4ade4dc816fa1700626d6fdccf62c599abfd91c62c3c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.11.15

Release files / orca_ai-0.1.1-py3-none-any.whl

Download URL orca_ai-0.1.1-py3-none-any.whl
Size 131.9 kB
Tags Python 3
SHA-256 checksum
How to use checksums
b83d039d6e8fa156e8b841809799340cc788e3e4b75a15fc9656b5b60a126676
BLAKE2b-256 checksum
How to use checksums
176f8d76316724a83b1e5a2cc7b4ed1010ec229f00eb2587471ac10d1f21195f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.11.15

Release history Release notifications | RSS feed

This release

0.1.1 This release

2 release files

0.1.0

2 release files

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