Is there a community edition of Palantir? Meet OpenPlanter: An Open Source Recursive AI Agent for Your Micro-Surveillance Use Cases

The balance of power is changing in the digital age. While governments and large corporations have long used data to track individuals, a new open-source project called openplanter That power is being given back to the people. Created by a developer’Shin Megami Boson‘,OpenPlanter is a recursive-language-model checking agent. Its goal is simple: Help you keep an eye on your government, because they’re almost certainly keeping an eye on you.

Solution to the ‘odd data’ problem

The investigation task is difficult because the data is messy. Public records are often spread 100 Various formats. you probably have one csv Of campaign finance records, a JSON file of government contracts, and a PDF Disclosure of lobbying.

OpenPlanter swallows them Different structured and unstructured data sources With ease. It uses large language models (LLM) to perform unit resolution. This is the process of identifying when different records refer to the same person or company. Once it connects these points, the agent potentially looks for anomalies. It looks for patterns that a human might miss, such as a sudden increase in contract wins after a specific lobbying event.

Architecture: Recursive Sub-Agent Delegation

What makes OpenPlanter unique is its recirculating engine. Most AI agents handle 1 Make requests at one time. However, OpenPlanter breaks larger objectives into smaller pieces. If you give it a big task, it uses sub-agent delegation strategy.

agent defaults max-depth 4. This means that the main agent can give rise to a sub-agent, which can give rise to another, and so on. These agents work in parallel:

1. resolve entities In large scale datasets.
2. link dataset Those who do not have a common ID number.
3. Create evidence chains Which supports every single search.

This recursive approach allows the system to handle investigations that are too large for a single ‘context window’.

2026 AI Stack

OpenPlanter is designed for high-performance requirements 2026. it is written in Python 3.10+ And integrates with the most advanced models available today. The technical documentation lists several supported providers:

• OpenAI: it uses GPT-5.2 As default.
• anthropic:it supports cloud-opus-4-6.
• openrouter:this is the default Anthropic/Cloud-Sonnet-4-5.
• cerebrus: it uses quen-3-235b-a22b-instructions-2507 For high speed operations.

also uses the system Example For web searches and voyage For high-accuracy embeddings. This multi-modal strategy ensures that the agent uses the best ‘brain’ for each specific sub-task.

19 tools for digital forensics

agent is equipped 19 Special equipment. These tools allow it to interact with the real world instead of just ‘chatting’. These are organized into 4 main areas:

• File I/O and Workspace:equipment like read_file, write_fileAnd hashline_edit Allow the agent to manage its own database of findings.
• shell execution:agent can use run_shell To execute the actual code. It can write a Python script to analyze the dataset and then run that script to get the results.
• web recovery: with web_search And fetch_urlIt can pull live data from government registries or news sites.
• planning and reasoning: The think The tool allows agents to pause and strategize. it uses acceptance criteria To verify that a sub-task was completed correctly before moving on to the next step.

deployment and interface

OpenPlanter is designed to be accessible yet powerful. It has a feature Terminal User Interface (TUI) made with rich And prompt_toolkit. The interface consists of a splash art screen of ASCII potted plants, but the work it does is serious.

You can start using it quickly postal worker. by driving docker compose upThe agent starts in a container. This is an important security feature because it isolates the agent run_shell Commands from the user’s host operating system.

Command-line interface allows ‘headless’ tasks. You can run a single command like:

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

The agent will work autonomously until the final report is generated.

key takeaways

• Autonomous Recursive Logic: Unlike standard agents, OpenPlanter uses Recursive Sub-Agent Delegation Strategy (default max-depth 4). It breaks complex investigative objectives into smaller subtasks, parallelizing work among multiple agents to create detailed evidence chains.
• Anomalous Data Correlation: The agent is designed to be swallowed and dissolved Separate structured and unstructured data. It can process simultaneously csv files, JSON records, and unstructured text To identify entities in segmented datasets (like PDF).
• Potential anomaly detection: by performing unit resolutionOpenPlanter automatically adds records – such as matching corporate surname to lobbying disclosure – and searches. potential anomalies To expose the hidden connections between government spending and private interests.
• High-end 2026 model stack: The system is provider-agnostic and uses the latest Frontier models, including OpenAI GPT-5.2, Anthropic Cloud-Opus-4-6And Cerebras Quen-3-235b-a22b-instructions-2507 For high speed estimation.
• Integrated toolset for forensics: OpenPlanter Features 19 specific equipment including shell execution(run_shell), Web Search (EXA)And file patching (hashline_edit). This allows it to write and run its own analysis scripts, validating the results against real-world acceptance criteria.

check it out Repo here. Also, feel free to follow us Twitter And don’t forget to join us 100k+ ml subreddit and subscribe our newsletter. wait! Are you on Telegram? Now you can also connect with us on Telegram.

Disclaimer: MarkTechPost does not endorse the OpenPlanter project and provides this technical report for informational purposes only.