Project Jupyter

Panel is a research workspace that places an AI-agent chat alongside files, PDFs, and live Jupyter notebooks. The agent can read and write files, run notebooks against a real kernel, create scratch data, and edit the same notebook as the user; configurable panes can also display images, data, code, chats, and custom viewers or apps. It supports workspaces with separate chats and saved layouts, background commands, and literature reviews. The early build currently has full support for Claude Code, with optional OpenAI API support for chat and tools; it is distributed under the MIT license and runs locally with Node, pnpm, uv, and Python.

CodeJury is a terminal-first, knowledge-grounded multi-agent software delivery pipeline for scoping requirements, implementing changes, running tests, and gating pull requests. It divides delivery into six terminal-driven stages and uses isolated development branches, human approval gates, deterministic QA, and ensemble code review. The system indexes a repository into a persistent code graph and uses semantic search to retrieve relevant code during later work. Its review stage sends the same change to independent models from different providers; both must approve it for the change to pass, while a rejection returns the reason to the developer. A configurable panel can use up to six specialist reviewers and a foreperson to reconcile their findings. The project is distributed as a Python package installable with pip and documents support for macOS, Linux, and Windows. Its repository describes a one-command run and operation without an API key.

Jive is an open-source terminal coding agent that plans work as executable graphs. It replaces sequential tool calls with graph calls: the planner produces a directed acyclic graph containing tool calls and Jev calls, allowing the agent to perform multi-step profiling, dataset analysis, and repetitive tasks with fast decisions inserted between harder reasoning steps. The project emphasizes a lightweight, token-efficient agent scaffold with eager execution, flexible graphs, and support for system-one and system-two models. It can be installed with the project's shell installer and is distributed under the MIT license.

Clay Studio is a self-hosted, browser-based team workspace for running and sharing local coding-agent sessions across projects. It connects to supported local runtimes such as Claude Code and Codex, allowing people and AI collaborators to share live sessions, hand off work, respond to permission requests, and retain project knowledge between conversations. The Clay Studio daemon runs on the user's machine and serves the workspace to a browser or mobile PWA. It stores sessions and knowledge on disk as portable JSONL and Markdown, and supports repositories, parallel sessions, Git worktrees, background workers, scheduled tasks, persistent AI collaborators called Mates, project instructions, file editing, diffs, terminals, MCP servers, and push notifications. The request path is browser to the Clay Studio daemon to the selected coding-agent runtime and its model provider; project traffic is not relayed through a Clay-hosted cloud.

PI-Desktop is a local-first desktop workspace for AI coding agents, built with Electron, a Rust host core, and the pi Agent Harness. It lets users open local projects, connect cloud or local models through providers such as OpenAI-compatible APIs, manage projects and long-running sessions, and review file changes and command output. Its architecture separates the React renderer, Electron desktop orchestration, Rust-managed permissions, filesystem access, SQLite persistence and secrets, and a pi Agent sidecar responsible for the agent loop, model interaction, and streaming. Agent, Plan, and Goal modes provide different approval boundaries; agents can read and edit files, run commands, and delegate exploration, implementation, research, testing, or review to background Subagents. The workspace supports MCP servers, Skills, installable Plugins, model and provider switching, session imports, notifications, context checkpoints, and a plugin marketplace. Conversations are stored locally as JSONL with a SQLite index, settings and logs remain on the machine, credentials use the operating-system keychain, and model requests go directly to the configured provider or endpoint. Plugin processes are permission-gated and isolated from the renderer, but plugins remain user-trusted code rather than a complete operating-system sandbox. It is an early-preview cross-platform application distributed for macOS, Windows, and Linux under the GNU Lesser General Public License v3.0.