OpenPlanter: A Recursive Open-Source AI Agent for Micro Surveillance and Data Investigation

Is There a Community Edition of Palantir? Meet OpenPlanter: An Open Source Recursive AI Agent for Your Micro Surveillance Use Cases

OpenPlanter, developed by ‘Shin Megami Boson’, is a recursive-language-model investigation agent designed to enable public oversight of government activities. The system utilizes a unique recursive engine with a default max-depth of 4 to break complex investigative objectives into parallelized sub-tasks.

Why This Matters

Investigative work often fails due to the ‘Heterogeneous Data’ problem, where public records are scattered across hundreds of incompatible formats like CSV, JSON, and PDF. While ideal models assume clean data, OpenPlanter addresses technical reality by using LLMs for entity resolution and probabilistic anomaly detection to connect records without common ID numbers. This approach allows citizens to surface hidden connections between government spending and private interests that traditional manual analysis would miss.

Key Insights

• Recursive sub-agent delegation allows OpenPlanter to handle investigations larger than a single context window by spawning agents up to a depth of 4.
• Entity resolution via LLMs identifies matching individuals or companies across disparate structured and unstructured formats like CSV, JSON, and PDF (2026).
• The system utilizes a high-performance 2026 model stack including OpenAI gpt-5.2 and Cerebras qwen-3-235b-a22b-instruct-2507 for high-speed inference.
• Digital forensics is enabled through 19 specialized tools, including shell execution via run_shell and automated file patching via hashline_edit.
• Docker-based deployment isolates the agent’s shell commands from the host OS, providing a critical security layer for autonomous code execution during analysis.

Working Examples

Command-line interface execution for a headless autonomous investigative task.

openplanter-agent --task "Flag all vendor overlaps in lobbying data" --workspace ./data

Practical Applications

• Public Oversight: Identifying vendor overlaps in lobbying data by correlating campaign finance records with government contracts.
• Digital Forensics: Executing Python scripts within a container to analyze large datasets and construct evidence chains automatically.
• Anti-pattern: Running run_shell commands without Docker isolation can expose the host operating system to unintended script behaviors.

References:

• https://www.marktechpost.com/2026/02/21/is-there-a-community-edition-of-palantir-meet-openplanter-an-open-source-recursive-ai-agent-for-your-micro-surveillance-use-cases/