AIAny
AI Agent2026
Icon for item

nodeterm

Manages real tmux-backed terminals and AI agents as draggable nodes on an infinite pan/zoom canvas, with a Trello-style kanban view, persistent sessions that survive restarts, mobile companion support, and a browser Server Edition for self-hosting.

Introduction

Most developer UIs treat terminals and agent sessions as ephemeral tabs or panes, which buries long-running work and breaks spatial context. nodeterm flips that model: terminals and agents are first-class, persistent nodes on an infinite canvas so you can arrange, group, and visually track parallel workstreams across restarts and devices.

What Sets It Apart
  • Persistent tmux-backed sessions: nodes map to real tmux sessions that survive app restarts and machine reboots, so long-running processes and scrollback are preserved instead of lost in transient tabs. This reduces accidental context loss for background tasks and agents.
  • Agent-aware UX and orchestration hooks: agents (Claude Code, Codex, Gemini, custom CLIs) surface RUNNING / NEEDS YOU states, permission prompts, and live transcripts; agents can be linked as context and even drive canvas actions via a CLI, turning the canvas into a lightweight mission control for multi-agent workflows.
  • Two-surface model (canvas + kanban) and mobile continuity: every project is both an infinite canvas and a Trello-style board of live sessions, and a paired iOS app or Server Edition lets the exact same live tmux sessions be accessed from a phone or browser with end-to-end encrypted relay, enabling on-the-go monitoring and input.
  • Remote-first transports and source-control integration: Local and remote (SSH) projects present the same canvas UI, with git worktree binding, file/editor panels, and GitHub issue cards mapped to the kanban — useful when agents or builds must run on remote hosts but you want a single local control surface.
Who It's For and Trade-offs

Great fit if you juggle multiple long-running shells or AI agents across repos and machines and prefer spatial organization over stacked tabs — e.g., researchers running parallel experiments, devs coordinating agent-backed tasks, or engineers monitoring remote CI/agents. It’s especially useful when visual grouping, session continuity, and mobile/remote oversight matter.

Look elsewhere if you need a minimal terminal emulator or a lightweight tabbed terminal: nodeterm is an Electron app with additional overhead and opinionated features (canvas, kanban, built-in tmux). Also, while the Server Edition enables remote access, self-hosting introduces the usual operational considerations (service management, updates, and secrets handling). The project uses a BUSL-1.1 license (be mindful of the Additional Use Grant if you plan to offer a competing hosted service).

More Items

Hugging Face

Provides 5.5K+ self-contained data-analysis RL tasks: each row bundles a real tabular dataset, a question, and a deterministically-gradable gold answer. Verified from jupyter-agent notebooks; splits for training, held-out testing, and quick eval; intended for prompting, fine-tuning, and agent RL.

Hugging Face
AI Model2026

A 9B agentic multimodal SFT checkpoint distilled from Qwen3.5-9B for coding, general agent tasks, visual coding and cybersecurity. Provided by Xiaomi MiMo as a research seed (77.4B-token SFT mix) to bootstrap agentic RL and tool-use experiments.

Hugging Face
AI Model2026

Preview agentic language model for research and engineering workflows that turns research questions into executable, verifiable workflows via tool use and long-context reasoning; built on a 744B-parameter MoE (GLM-5.2) with MIT-licensed BF16 and FP8 checkpoints.