Hivekeep
VerifiedSelf-hosted persistent AI agents that collaborate and remember on your server.
What is Hivekeep?
Hivekeep is an open-source self-hosted platform that deploys multiple persistent AI agents inside one Docker container. These agents maintain continuous memory across sessions, collaborate by requesting and replying to each other, and can create their own plugins or mini-apps over time.
The system uses a single SQLite file for storage and an encrypted vault for secrets that never reaches the models. Agents connect to users through Telegram, WhatsApp, Slack and other channels while supporting scheduled work and real-time handoffs between specialists.
It targets users who want reliable, private AI assistance without relying on cloud services or losing context between conversations.
What you can build with Hivekeep
Household task coordination
Agents track reminders, calendars, and backups in one ongoing conversation while delegating subtasks among themselves.
Personal project management
A shared kanban board lets agents create tickets, update progress, and build supporting mini-apps directly inside the workspace.
Multi-channel personal assistant
Users interact via any connected messaging app while agents maintain context and hand conversations to specialized agents as needed.
Install Hivekeep
docker run -v ./data:/data ghcr.io/marlburrow/hivekeepcurl -fsSL https://raw.githubusercontent.com/MarlBurroW/hivekeep/main/install.sh | bash- 1Pull the official Docker image from ghcr.io.
- 2Run the container with a mounted volume for the SQLite database.
- 3Access the web interface or connect a messaging channel.
- 4Start a conversation with the default agent Queenie to configure the team.
- 5Let agents create initial tools and plugins through natural language requests.
Hivekeep: pros & cons
Pros
- +True persistence with hybrid long-term memory that never resets.
- +Single-container deployment with no external databases or brokers required.
- +Agents can build and host their own tools and mini-apps.
- +Strong privacy through local secrets vault and full data ownership.
Cons
- –Requires self-hosting and basic Docker knowledge to get started.
- –All compute and storage stays on your hardware, limiting scale.
- –Multi-agent coordination may need initial setup conversations.
Frequently asked questions
Yes, it is released under the MIT license with no paid tiers.
User reviews
Verified reviews from the community shape this listing's rating.
Loading reviews…
Similar agents
Other general-purpose options worth comparing.
