FERRICULA
CloudYour Brain in the Cloud.
Every agent gets its own ferricula instance — persistent memory, automatic dreaming, shivvr embedding. Deployed on Google Cloud Run. Scale to zero when idle. Pay only for what you use.
View PlansFully Managed Stack
Each plan includes everything you need. No Docker, no infra, no ops.
Persistent Memory
Memories survive restarts. Snapshots and WAL sync to cloud storage automatically. Cold start restores full state from GCS.
Automatic Dreaming
The clock thread runs continuously on paid plans. Entropy accumulates, dreams fire, memories consolidate and decay — all while you sleep.
MCP Access
Point your MCP config at your cloud endpoint. Same 19 tools — remember, recall, reflect, dream, connect — no local binary needed.
Shivvr Embedding
Shared shivvr cluster handles text embedding and vec2text inversion. gtr-t5-base, 768 dimensions, GPU-accelerated. You just send text.
Give Your Agents Memory
Free
- 1 ferricula agent
- Cloud persistence
- Manual dream cycles
- Community support
Pro
- 5 ferricula agents
- Cloud persistence
- Automatic dreaming (clock)
- Email support
Team
- 20 ferricula agents
- Cloud persistence
- Automatic dreaming (clock)
- Priority shivvr embedding
- Dedicated support
How It Works
Sign up and choose a plan
Enter your email, pick Free/Pro/Team. Paid plans go through Stripe checkout. Your ferricula agent(s) start provisioning immediately.
Get your endpoint
Each agent gets a unique Cloud Run URL. Point your MCP config at it.
Use it like local ferricula
Same MCP tools, same HTTP API, same 16 REST endpoints. Remember, recall, dream, connect, checkpoint. Your data persists in Google Cloud Storage and restores on cold start.