Voidleap Code é um IDE para desktop voltado a desenvolvedores profissionais e engenheiros de IA que criam software com agentes de IA. Ele é executado localmente no Mac e no Windows, oferece suporte a contas de provedores de nuvem ou modelos locais e permite inspecionar e controlar prompts, chamadas de ferramentas, contexto e alterações em arquivos.
Voidleap Code is a desktop integrated development environment for building software with AI coding agents. It combines a code editor with an agent harness that lets developers assign tasks, inspect prompts and tool calls, review file changes, stop or correct an agent, and fork threads for alternative approaches.
The application runs locally on the developer’s computer and supports user-supplied cloud provider accounts or local models. Its stated architecture keeps application data on disk; when cloud inference is used, code and prompts pass directly between the local app and the selected provider rather than through Voidleap’s servers.
The desktop app runs locally and can use supported cloud providers, Ollama, or an OpenAI-compatible local server. Agents, skills, and workflows can be used across providers, with provider changes possible during a thread.
Prompts, model calls, tool calls, messages, and file changes are visible. Users can edit context, remove items, stop an agent, fork a thread, and run multiple agents in parallel panels.
Four execution modes range from read-only to fully automatic operation. Bash and PowerShell commands are parsed before execution, output streams live, sensitive paths are blocked by default, and an optional OS sandbox supports network allow and deny lists.
Agents, skills, hooks, MCP servers, slash commands, keybindings, security rules, and automations can be edited visually or as files. Global, workspace, and project scopes support shared configuration with local overrides.
The dashboard reports cost, tokens, latency, and cache hits, with a live model-call feed. Users can trim or hand off context and control which tools and skills enter the system prompt.
A built-in browser supports annotations for agents, and a local code index supports navigation. The marketplace and bundle editor provide plugin creation, review, publishing, and per-scope installation workflows.
Assign an implementation or maintenance task, monitor commands and file edits, intervene when needed, and fork the thread to compare another approach.
Connect provider accounts, change models during a thread, or combine providers when different parts of a task call for different model capabilities.
Configure Ollama or another OpenAI-compatible local server for work that should remain on local hardware and avoid provider transmission.
Package recurring instructions into agents, skills, hooks, slash commands, MCP connections, or automations, then share them through global, workspace, or project scopes.
The planned Team and Enterprise tiers target shared servers, scoped access, collaboration, SSO, central policy, and on-premises deployment; these capabilities are not described as fully available today.
The desktop app runs on the developer’s computer. The site identifies Mac and Windows as supported today, with a Unix desktop app in internal testing. Application data is stored locally, including under ~/.voidleap/ according to the documentation.
No. Users connect their own provider subscription or API key, or configure a local model through Ollama or an OpenAI-compatible server. Provider charges and terms apply separately.
The product states that code and prompts do not pass through Voidleap’s servers. They move between the local application and the provider selected by the user; Voidleap’s servers are used for license checks and a periodic model-pricing configuration fetch.
Yes, depending on the selected execution mode. The product offers four modes, command parsing, live output, blocked sensitive paths, and an optional OS sandbox. Users can require approval or allow more automatic execution.
The Free version is available now. Pro and Team are described as in development; planned Team capabilities include shared servers, scoped invitations, real-time collaboration, handover, and managed seats. Enterprise is planned to add on-premises deployment, central policy, SSO, and key pooling.
Os dados de tráfego são apenas para referência.
kilo.ai
Kilo Code é um plugin de programação com IA multiagente para JetBrains IDEs. Desenvolvedores podem executar e gerenciar agentes de programação com IA, revisar alterações em worktrees isolados e escolher entre um amplo catálogo de modelos sem sair do IDE.
caveman.so
Caveman é uma stack de eficiência de tokens para agentes de IA e fluxos de trabalho em produção. Ajuda as equipes a reduzir o contexto elegível, rotear o tráfego e medir economias verificadas com ferramentas locais, um SDK TypeScript e uma camada de nuvem gerenciada.
apps.mergeable.io
Diffsmith é um app nativo para macOS que permite revisar alterações locais de git feitas por agentes de programação com IA, como Claude Code, Cursor, Codex e Copilot. Adicione comentários em linha e envie-os ao agente como prompt ou por meio de um servidor MCP local.
www.clarkchat.com
Clark Labs oferece Clark Agent para trabalho web autônomo baseado na nuvem e Clark Code para tarefas de programação locais ou baseadas em SSH. A plataforma também fornece um gateway de API compatível com OpenAI para chamar modelos Clark ou encaminhar modelos de ponta por meio do Clark.
harnessrouter.ai
HarnessRouter é uma camada de backend para agentes que permite às equipes integrar Codex, Claude Code e Hermes ao produto por meio de uma única API, na própria infraestrutura. Também oferece uma opção de nuvem gerenciada para equipes que preferem não operar o runtime dos agentes por conta própria.
agentgrid.sh
AgentGrid é um aplicativo para desktop que gerencia vários agentes de IA em uma única tela infinita, reunindo terminais, navegadores, notas e ferramentas de código no mesmo espaço de trabalho. Foi criado para desenvolvedores que desejam orquestrar visualmente o trabalho paralelo de agentes no macOS, Windows e Linux.
