Neural Inverse Build / Editor
Write, review and verify code with AI, under your organisation's policy.
A desktop development environment built on the open-source core of Visual Studio Code, for engineers in regulated work.
- Part of
- Neural Inverse Build
- Platforms
- macOS, Windows and Linux
- Built on
- Code - OSS
- Models
- Through Govern, your own provider or local
- Record
- Tamper-evident ledger on the device
Overview
The Editor is where engineers do hands-on work with AI: completions while they type, questions in chat, and edits they review line by line. Every change is checked, every governed action is decided by your organisation's signed policy, and the record stays on the device unless your organisation decides otherwise.
What it does
What the Editor does.
What an engineer needs to work with AI, and what a reviewer needs to trust the result.
AI assistanceInline completions while you type, questions in chat, and proposed edits shown as changes for you to review line by line before you accept them.
ChecksDeterministic analyzers run on the changes you and agents make: structural rules for C and C++, Python, JavaScript and TypeScript, and secret detection on every file. The Editor, the CLI and CI produce the same findings.
Only signed rules blockOnly rules in your organisation's signed policy can block a change. Rules shipped with the Editor or found in a workspace are advisory. Nothing a model produces is a finding, and nothing it produces can block.
Activity LogWhat people, agents, extensions and command-line tools do is recorded in a local, encrypted, tamper-evident ledger you can read in the Activity Log.
Evidence packsA signed pack for a release, a change, a trace or a time range: which lines AI touched, which model and context were used, who approved, what left the boundary. Anyone can verify a pack offline.
Domain modulesFirmware, FPGA and HDL, PLC, DCS, mainframe, substation and automotive work, driving the vendor tools you already use. Each module's verification state is published.
ModelsYour organisation's models through Govern with your own sign-in, a provider you connect with your own key where policy allows, or a model on your own machine.
Shared with the Agents appThe Editor's agent features run in the same agent host as the Agents app, so both see the same model providers and credentials.
Key workflow
From sign-in to a signed evidence pack.
- InstallYour administrator provides the installer and the address of your Govern deployment.
- Sign inIn the browser or with a device code. The status bar shows your connection.
- Choose a modelYour organisation's models, or a provider or local model where policy allows.
- Work with AICompletions, chat and proposed edits you review line by line.
- Check and approveChanges are checked. Actions your policy marks wait for a person.
- EvidenceBuild and verify a signed evidence pack for the release or change.
Who uses it
Built for every role on the team.
Software engineersWrite and review code with AI, with checks on C and C++, Python, JavaScript and TypeScript changes.
Firmware engineersImport Keil, IAR and STM32CubeIDE projects into CMake, build with a size report, flash after a confirmation and debug without a launch file.
FPGA and controls engineersLint, simulate and read timing reports beside the HDL; open TwinCAT and DeltaV work in the Editor. Vendor paths are in verification.
Reviewers and quality teamsRead the Activity Log and verify evidence packs offline, with every gap listed.
How it is governed
Governed on the device, enforced where it binds.
The Editor applies your organisation's policy locally. What must hold is enforced by Govern and your CI.
Signed policyYour organisation's policy arrives signed and is verified on the device before any of it is used. An older policy is refused, and deleting the cached one locks rather than loosens.
ApprovalsWhen policy says an action needs approval, you are asked before it runs. Accepting agent changes, installing extensions, running tasks, changing workspace trust, adding MCP servers, changing settings and choosing models are decided by one governance service.
Built-in protectionWith or without an organisation policy, an agent cannot change policy and approval records, git hooks, CI configuration, shell profiles or audit data without a person.
Locks when policy lapsesWhen a policy has expired past its offline grace, or a required policy cannot be verified, agent sessions, tool calls and model calls lock. Editing and saving files keep working.
Audit set by managed policyOnly managed policy sets what the ledger records; a settings file cannot turn audit off. By default nothing leaves the device. Your organisation can turn on uploading to its own Govern deployment.
What bindsA modified build can remove any check inside the Editor, so model access is enforced by Govern's gateway, and merges and releases by your CI.
Technical facts
For your IT and security review.
PlatformsmacOS, Windows and Linux. Installed together with the Agents app; each has its own process, data folder and sign-in session.
Sign-inYour organisation account on your organisation's Govern deployment: in the browser with OAuth 2.0 and PKCE, or with a device code (RFC 8628) on a machine with no browser. The session is kept in the operating system's credential store.
Deployment checkBefore sign-in the app verifies the deployment's region document and issuer over HTTPS and pins the issuer. Tokens are only sent to the deployment that issued them.
Managed policyGroup Policy templates on Windows, a configuration profile on macOS and a JSON policy file on Linux. The Govern deployment address and audit settings are set this way.
ModelsGovern's model gateway, compatible with the OpenAI and Anthropic APIs; Anthropic, OpenAI, Google Gemini, Azure OpenAI, Vertex AI, Bedrock and other providers with your own key; Ollama, LM Studio and vLLM locally.
Without internet accessLocal models need no key, Govern can be self-hosted on premises, and policy is decided on the device with no Neural Inverse server called. No telemetry is sent to Neural Inverse.
Licence and statusProprietary, under your agreement with Neural Inverse; source review under NDA. Pre-release: interfaces and data formats may change between versions.
Get started
Install, sign in and read the documentation.
Talk to our team.
Neural Inverse is available by request. We work with your engineering, security and procurement teams from the first conversation.