The Bitter Lesson applies to the harness, too.
The model should decide what to remember, where to look, when to delegate, and how to proceed. Plurnk is an agentic operating system and programming language that puts those decisions in the model's hands.
Files, tools, and the agent's own context become an addressable environment. Composable operations let it search that environment, make precise changes, curate its memory, and build its own delegation topology. The runtime provides reliable machinery; the model supplies the strategy.
Use local or cloud models to build software, investigate a codebase, or automate work across tools. Interact from your terminal or editor, or compose Plurnk with ordinary shell pipelines.
This repository provides the CLI and interactive terminal client for plurnk-service, the shared daemon.
- Curation, not compaction. The agent retrieves the passages it needs and removes stale items or individual lines from its active context. Its working set changes; source material and original execution evidence survive.
- Precision without ceremony. Line ranges, character regions, and hash-anchored edits make surgical changes possible. Stale anchors reject conflicting edits before they overwrite the wrong text.
- The model chooses the topology. Fork with existing context, start a worker with a fresh log, or delegate pure inference without an agent loop. Parent and child models can use different endpoints: a cloud model can orchestrate local workers through the same primitives.
- Execution with evidence. ANTLR parses model output into executable operations. The runtime records their results and returns structured errors the model can act on. Full packet digests make the work inspectable. A compatible local server can run under a grammar you wrote, carried verbatim.
Workspaces and worker conversations live in the daemon, independently of the client session. Choose your model, context limits, tools, and capability policies without replacing the environment. No Plurnk account is required.
The pattern engine connects discovery and context management. Path globs combine with full-text search, regex, JSONPath, XPath, and symbol-graph queries. The model can search for phrases, inspect structured data, or follow symbol relationships without writing a script for each question.
For example, these model-side operations find TypeScript files matching
retry or timeout, then trim older READ receipts to their first 16 lines:
### FIND_ (src/**/*.{ts,tsx})
~retry OR timeout
### KILL_ (log:///1/[1-7]/*/READ) <17,-1>
The second operation targets READ results from turns 1–7 of loop 1. It curates the log, not the source files. One expression can manage many entries: the agent has bulk operations over its own context, not just over your code.
Requires Node.js 26+, npm, Git, and a local or cloud model endpoint.
npm install -g @plurnk/plurnk @plurnk/plurnk-serviceIn one terminal, configure a model and start the daemon. This example uses DeepSeek; see model configuration for other providers and local servers.
export DEEPSEEK_API_KEY="your-api-key"
export PLURNK_MODEL=deepseek/deepseek-v4-flash
plurnk-service startIn another terminal, open a project:
cd /path/to/your/project
plurnk --workspace="myProject"Give it a task in ordinary language. The model uses the operation language; you do not need to learn it to use Plurnk. Run the same command later to return to that workspace's conversation.
The client connects to 127.0.0.1:1066 by default and never starts the daemon.
Provider credentials belong in the daemon's environment. Proposals are accepted
automatically by default (PLURNK_CLIENT_YOLO=1); set it to 0 or start a prompt
with ? to review them. Neither overrides capability restrictions.
plurnk --workspace="myProject"A scrollback-native TUI with multiline prompts, streaming reasoning when the
provider supplies it, Markdown and Mermaid rendering, and slash commands for
managing models, workers, and tools. Run /help to explore; /model and
/child select the conversation and delegated models.
plurnk "Explain how this project's request handling works"
plurnk "Summarize this repository" > overview.md
git diff | plurnk "Review this patch for correctness"
plurnk --json "Explain the test layout" | jq -r .responseOne-shot commands put the answer on stdout and progress on stderr. --json
returns one structured document containing the answer, operation trace,
diagnostics, and usage.
Use plurnk --help for CLI options. plurnk models lists available model
routes.
For an editor-native interface, use plurnk.nvim against the same daemon.
Plurnk uses AG-UI for clients, MCP for external tools, Agent Skills for reusable instructions, and A2A for remote agents. Tools, skills, and agents can be discovered and enabled per worker without restarting the daemon. Their documentation is retrieved on demand rather than loading every tool's schema into every prompt.
Configuration uses cascading environment variables and .env files, including
the XDG user configuration at ~/.config/plurnk/.env. Model discovery uses
models.dev. Inspect the complete, documented configuration
catalog with:
plurnk-service config defaults- Installation and configuration
- Architecture and extension points
- The model-facing language
- Client reference and terminal design
Questions, bug reports, and feedback are welcome in GitHub issues. See the shared contributing guide for development guidance.
MIT.