Neural Inverse IDE
An AI-native development environment for engineers writing software where correctness matters. Runs on your machine. Free under Apache 2.0. No account required.
What Is Included
The full IDE ships as a single install. Everything below is in the open-source release with no usage limits and no sign-in required.
AI Coding - inline chat (Ctrl+L), inline edit (Ctrl+K), fill-in-the-middle autocomplete, and AI code actions. Works with any LLM provider you configure.
Power Mode - agentic coding environment with a terminal-style interface. Reads files, runs commands, edits code, checks output, iterates. Runs sub-agents scoped to specific roles: explorer reads, editor modifies, verifier tests. Open with Cmd+Alt+P.
Firmware Development - hardware-accurate AI assistance for embedded systems. 357 MCU variants across 11 manufacturers. SVD register maps, datasheet intelligence, MISRA C, IEC 62304, ISO 26262, DO-178C compliance support, GDB debug agent, RTOS debugger, logic analyzer, RTT viewer, clock tree, pin mux. Open with Cmd+Alt+F.
Legacy Modernisation - AI-assisted migration for legacy codebases. 25+ migration patterns including COBOL, RPG, Oracle, monolith decomposition, and framework upgrades. AI-generated migration plan, two-window source/target editing, compliance fingerprint comparison. Open with Cmd+Alt+M.
Agent Manager - manage agentic workflows, configure sub-agent roles, set tool permissions per agent. Open with Cmd+Alt+A.
20 LLM Providers - Anthropic, OpenAI, Azure OpenAI, Google Gemini, AWS Bedrock, Ollama, LM Studio, and more. Bring your own key. Your code never leaves your machine unless you send it to the model.
Install
# macOS / Linux
curl -fsSL https://cdn.neuralinverse.io/ce/install.sh | bashOr download a release directly from GitHub for macOS, Linux, and Windows.
Data Sovereignty and Privacy
The local IDE is the only AI coding environment where your code never leaves your machine unless you explicitly choose to send it.
Data sovereignty - your source code, your configuration, your LLM prompts, and your completions all stay on your machine. There is no cloud service in the request path. You decide what goes where.
No vendor lock-in - Neural Inverse does not own your model access. You bring your own API keys for any of the 20 supported providers. You can switch providers mid-session. If a provider changes pricing or shuts down, you change a key and keep working.
Privacy by default - there is no telemetry that leaves your machine without your consent. No user analytics, no prompt logging, no training data collection from your code. The IDE is a local application. What you write stays with you.
Air-gapped and regulated environments - combine the local IDE with a self-hosted or Ollama model and you have a fully offline AI development environment. Nothing crosses the network boundary.
License
Apache 2.0. Use it commercially, modify it, fork it, self-host it.
What Cloud and Enterprise Add
The local IDE is the foundation. Two hosted tiers build on top of it:
Neural Inverse Cloud - managed workspaces running the same IDE in a browser tab. Private git storage, free OSS models, on-demand hardware simulation. No local install needed. Learn more
Neural Inverse Enterprise - adds the governance layer for regulated organisations. Real-time GRC compliance enforcement, IAM policy engine, Enclave, AI governance controls, and an immutable audit trail. Runs locally with your team or hosted on platform.neuralinverse.com. Learn more