One goal,
many experts.

Define a team of specialist agents. Talk to the orchestrator. Planck handles the rest.

01

Define your team

Planner, builder, reviewer — or whatever the workflow demands. Declare roles, prompts, and tools in a TEAM.json.

02

Agents coordinate

Each agent runs as an OTP process. Delegation, responses, and interrupts flow as BEAM messages — no queues, no glue code.

03

You talk to the orchestrator

One conversation. The orchestrator breaks the goal into tasks, delegates to specialists, and reports back when done.

# Install curl -fsSL https://thebroken.link/planck/install.sh | sh # Run planck # Open http://localhost:4000

On macOS, if Gatekeeper blocks the binary, the install script removes the quarantine flag automatically. If you downloaded manually: xattr -d com.apple.quarantine /path/to/planck

# Install (PowerShell) irm https://thebroken.link/planck/install.ps1 | iex # Run planck # Open http://localhost:4000
Documentation Hex.pm packages →

Run the full Planck stack on your own hardware. One command installs a complete environment: private web search, workspace indexing, document extraction, long-term memory, a credential proxy, and more. Requires Docker.

# Install (Linux / macOS) curl -fsSL https://thebroken.link/planck/install_docker.sh | sh # Open http://localhost:4000
# Install (PowerShell) irm https://thebroken.link/planck/install_docker.ps1 | iex # Open http://localhost:4000

Binds to 127.0.0.1:4000 by default. Pass --bind=0.0.0.0 (Linux/macOS) or -Bind 0.0.0.0 (Windows) to expose on the network.

SEARCH

Searxng

Private meta-search. Web results without sending queries to third parties.

INDEX

Typesense

Full-text search over your workspace files. Agents search before they fetch. Session turns are indexed for long-term memory.

DOCS

Apache Tika

PDF, DOCX, XLSX, PPTX and more — extracted to plain text for agents to read.

PROXY

Agent-vault

HTTPS credential proxy. API keys are injected into outbound requests by host — agents never see secrets, even via the bash tool.

MEMORY

Long-term memory

Past session turns are retrieved on resume and injected into context. The agent remembers previous conversations.

FOCUS

Short-term memory

Key facts and summaries are kept in a per-agent store that survives compaction and is injected at the start of every turn.

FETCH

Web fetch

Pages are fetched and converted to clean Markdown before being handed to the agent. Results are cached in Typesense — repeated fetches within a session are instant.

SKILLS

Skill reflector

After each turn the sidecar captures reusable patterns as skills. The agent becomes progressively more efficient over time.