OpenPlanter is a recursive language model agent designed for autonomous investigation of heterogeneous datasets. It targets engineers, researchers, and power users by ingesting diverse data sources, resolving entities across them, and surfacing non-obvious connections through evidence-backed analysis, thereby accelerating complex data exploration.

How It Works

The agent employs a recursive, sub-agent delegation architecture to tackle complex investigations. It autonomously operates using a suite of tools for file I/O, shell execution, and web search. Core to its design is the ability to ingest and link entities from disparate datasets like corporate registries and financial disclosures. Its recursive nature allows for parallelized entity resolution and evidence-chain construction, enabling deep dives into large datasets.

Quick Start & Requirements

Installation is straightforward via pip install -e . for editable mode or using docker compose up. Configuration requires API keys for supported LLM providers (OpenAI, Anthropic, OpenRouter, Cerebras) and services like Exa (web search), which can be set via CLI flags, environment variables, or .env files. The project requires Python 3.10+. Users can launch an interactive Terminal UI (openplanter-agent --workspace ...) or run single tasks headlessly (openplanter-agent --task ...).

Highlighted Details

• Features 19 distinct tools organized for dataset ingestion, shell execution, web search, and planning/delegation.
• Supports multiple LLM providers and models, allowing flexibility in processing.
• Recursive sub-agent delegation enables parallel investigation and deep analysis up to a configurable depth (max-depth).
• Offers both a rich Terminal UI and a headless mode suitable for CI/automation pipelines.

Maintenance & Community

The README does not detail specific contributors, sponsorships, or community channels like Discord or Slack. It directs users to VISION.md for the project's design philosophy and roadmap.

Licensing & Compatibility

The specific open-source license is not explicitly stated in the README. Users should consult VISION.md for licensing details. Compatibility for commercial use or integration with closed-source projects is not specified.

Limitations & Caveats

Operational costs are tied to external LLM API usage and web search services. The recursive nature, while powerful, can lead to extended execution times and significant costs if not carefully managed with depth and step limits. The accuracy of findings depends on data quality and LLM performance.