The Secure Orchestration Layer for BYOK Intelligence.Stop paying for three separate $20/mo subscriptions. Orchestrate GPT-5, Claude 4, and Llama 3 from a single, encrypted vault. Your keys, your memory, your budget.

Join the Beta Circle

Inside the circle

Key Vault: Encrypted API Key Storage — Your API keys are encrypted at rest and never leave your device. Only you hold the decryption key.

Memory Graph: Persistent Context Management — Visualize and manage your agent's persistent memory. Pin critical context, prune irrelevant data.

Token Budgets: Granular Cost Control — Set hard token limits per turn, per agent, or per session to control API costs and prevent context stuffing.

Agent Roles: Dynamic Model Orchestration — Define custom agent roles with preferred models, temperature, and specific system prompts for specialized tasks.

Multi-Model Orchestration: Seamless Provider Switching — Switch between GPT-4o, Claude 3.5, and Llama 3 on the fly. AICoven handles quota-aware fallback automatically.

Real-time Context Window Visualization — Monitor your context window in real-time. Understand token consumption and optimize prompt engineering.

Encrypted Chat History & Audit Trails — All interactions are end-to-end encrypted. Maintain a complete, auditable history of agent dialogues and tool calls.

Tool Orchestration: Dynamic Tool Loading — Agents dynamically load only relevant tools based on context, reducing token overhead and improving efficiency.

The Subscription Tax

Before AICoven

• Paying $720/year for redundant 'Pro' plans
• Fragmented AI chats across providers
• Lost memory between sessions
• Vendor lock-in

After AICoven

• Raw API costs (~$5/mo for high usage)
• Unified orchestration (quota-aware fallback)
• Persistent, searchable memory fabric
• Bring-Your-Own-Keys (BYOK) freedom

Technical Proof

role_definition.json

{
  "role": "Lead Architect",
  "primary_model": "claude-3-5-sonnet",
  "fallback": "gpt-4o",
  "context_window": "200k",
  "memory": "persistent",
  "encryption": "AES-256-GCM"
}

Define multiple agent roles with their own preferred model, autonomy level, and encrypted memory.

What makes it feel like magic

Unified Orchestration

Route prompts through any connected provider with quota-aware fallback.

Persistent Shared Memory

Recall your work across threads, roles, devices with approval workflows.

BYOK Security

Your keys never leave your vault. Encryption and audit trails by design.

Collaborative Covens

Shared workspaces with memory, roles, budgets, and transparent breadcrumbs.

Multiple Agent Roles

Define multiple agent roles with their own preferred model, autonomy level, max tokens, prompt.

Encrypted by Your Keys

Chats and memories are encrypted with your key. We do not store your chats in plain text.

Open Source

AICoven Local — Free & Open Source

Run AICoven entirely on your own machine. The local edition is fully open-source — no cloud account needed, no telemetry, no limits. Bring your own API keys and orchestrate every model with complete privacy and freedom.

View on GitHub

100% Local

All data stays on your machine. No cloud backend required.

Community Driven

Contribute features, report bugs, and shape the roadmap together.

All Features Unlocked

Every feature is available — covens, memory, multi-model orchestration.

Need help? Talk to the builders of the circle.

Questions about covens, memory, or BYOK routing? This form goes straight to the team behind AICoven's orchestration layer.

Most replies land within 24 hours on weekdays.