BeeSystemOpen Agent Platform

Self-hosting

Run the full BeeSystem stack on your own infrastructure. No data leaves your environment.

BeeSystem is designed to be fully self-hostable. All components — Bee Platform, Bee Worker, and Bee Hive — can run on your own infrastructure without any external dependencies.

Components

ComponentRequiredPurpose
Bee WorkerYesAgent execution runtime
Bee HiveNo (single node)Distributed orchestration
Bee PlatformNo (CLI/SDK)Agent design and management UI

Minimal self-hosted setup

# Pull images
docker pull beesystem/bee-worker:latest
docker pull beesystem/bee-hive:latest
docker pull beesystem/bee-platform:latest

# Start with Docker Compose
docker compose up -d

Data storage

BeeSystem does not require a specific database. The following backends are supported for persistent agent memory:

  • PostgreSQL (recommended for production)
  • SQLite (development and single-node)
  • Redis (fast session memory)
  • Any S3-compatible store (document and artifact storage)

Network requirements

Bee Worker and Bee Hive communicate on internal ports only. External network access is only required for configured tool connectors (e.g., APIs your agents call).

Security recommendations

  • Run Bee Worker and Bee Hive behind an internal load balancer
  • Enable auth on the Bee Worker API for production deployments
  • Restrict tool connector outbound access with firewall rules
  • Enable structured audit logging and ship to your SIEM