Skip to main content
2R

2389 Research

2389 Research builds on‑device AI agents that run locally on a user’s machine, accessing codebases, project files, and meeting transcripts to execute tasks without sending any data to external servers.

Chicago, IllinoisFounded 20235200+ followers
Updated 2 months ago

Funding

Funding not disclosed

Funding rounds are not available yet.

Founders

Product

Problem

Developers and knowledge workers often rely on cloud-based AI services that require sending proprietary code, ideas, and meeting data to external servers, creating privacy risks and dependence on third‑party infrastructure. This limits the ability to automate workflows locally and keep sensitive information confidential.

Solution

2389 Research provides on‑device AI agents that run entirely on the user’s machine, accessing local context such as codebases, project files, and meeting transcripts. The agents execute intentions—code refinement, workflow orchestration, and meeting summarization—without transmitting data externally, preserving privacy and reducing latency. A suite of tools (e.g., Simmer for iterative artifact refinement, dotpowers for DOT‑based development pipelines, Tracker for DAG workflow orchestration, and Summarize Meetings for knowledge‑graph generation) integrates with existing developer environments. The platform is open‑source where possible, allowing users to inspect and extend the underlying infrastructure while benefiting from research‑driven improvements.

Target Audience

Primary users are software developers, data scientists, and knowledge workers who need secure, local AI assistance for coding, automation, and information synthesis, as well as organizations that require on‑premise AI solutions to protect proprietary assets.

Features

  • On‑device execution of large language model agents, eliminating data upload to cloud services
  • Context‑aware operations that ingest local files, code repositories, and note vaults to perform tasks
  • Iterative refinement skill (Simmer) that scores and hones documents, prompts, or specifications over multiple rounds
  • DOT‑pipeline tooling (dotpowers, Tracker) for graph‑based workflow definition, human gating, and automatic checkpointing
  • Native macOS Graphviz editor (Dot Viewer) with live SVG preview for visualizing pipeline structures
  • Batch meeting transcript summarization that produces structured notes, participant tags, and knowledge‑graph links
  • Open‑source components and public research publications to foster transparency and community contributions
This profile is AI-generated and may contain inaccuracies.