Contact

Route the conversation by the action boundary.

Use the intake console to choose the right path: public Kernel review, Company AI OS architecture review, security disclosure, investment, press, or collaboration.

Start Technical Alignment View HELM AI Kernel
Primary path
Architecture review
Security
Dedicated route
Data
No secrets

Contact console

Ask Mindburn to route the work.

The assistant can route you toward HELM AI Kernel evaluation, HELM AI Individual access, Enterprise architecture review, Code Intelligence Graph review, autonomous business loop review, analog/kinetic simulator review, security disclosure, investor context, or technical collaboration using public sources.

Kernel evaluation Start from the public repo, docs, policy checks, and receipt model. Include: Include the action type, tool or connector, expected verdicts, and the proof artifact you need. Do not send: Do not send private keys, customer data, or internal infrastructure details. Architecture review Map the first governed action, authority boundary, connector, and receipt need. Include: Include the side effect, actor, approval rule, connector scope, escalation path, and review owner. Do not send: Do not attach credentials, raw logs with secrets, or regulated documents. Individual access Use this path for HELM AI Individual and one-person governed business loops. Include: Include the recurring business action, P0 ceiling, connector, approval rule, and receipt you need. Do not send: Do not send personal financial data, credentials, or private customer evidence. Code Intelligence review Use this path when engineering specs need code graph evidence, affected tests, and write-scope gates. Include: Include repository type, review goal, route/API risk, test expectations, and whether code comments may contain hostile text. Do not send: Do not send private source code, secrets, or repository credentials through the public assistant. Business loop review Use this path for Night Shift, support/refund, growth spend, finance/admin, or engineering-loop review. Include: Include allowed reads, writes, forbidden actions, approval thresholds, rollback, and closure evidence needs. Do not send: Do not send live customer records, bank details, or private work queues. Analog / kinetic simulator Use this path for simulator-labeled gateway review, not direct robot or factory control claims. Include: Include gateway type, safety profile, telemetry path, emergency-halt posture, approval rule, and jurisdiction boundary. Do not send: Do not send live operational credentials, site maps, or safety-critical private details. Design partner review Use this path when the Company AI OS layer needs workflow, approval, and integration context. Include: Include the workflow, current approval path, systems involved, and the receipt a reviewer would accept. Do not send: Do not send customer records, private evidence rooms, or unredacted policies. Security disclosure Use the right channel without sending secrets through the assistant. Include: Include affected public surface, reproduction steps, impact, and a safe contact path. Do not send: Do not exploit beyond proof, include secrets, or send customer data. Investor context Find the public company and product surfaces before a conversation. Include: Include your fund or program, thesis fit, timing, and the public surface you reviewed. Do not send: Do not ask for private customer proof or unpublished security evidence through public email.
Public source assistant

Mindburn contact assistant

Answers cite public sources and show the steps used.

answer ready with sourcescontext 6 / 7

Route a Company AI OS review request.

Start by naming the action boundary: what AI may propose, what must be approved, which connector would act, and what proof your team needs. [2][5]

Public evaluation can begin with HELM AI Kernel. Company AI OS conversations are reviewed access because they involve company workflows, integrations, approvals, and proof surfaces. [2][3]

Do not send secrets, credentials, customer data, regulated documents, or private evidence through this public assistant. Use the security channel for vulnerability reports. [5]

answer ready with sources5 sources
CompanyHomepageMindburn Labs builds HELM, the governed autonomous operations runtime for approved digital, analog, kinetic, and code workflows.HELMHELMModels propose. HELM governs execution and records proof before side effects leave the boundary.CompanyCompanyCompany mission, product map, operating principles, and founder archive boundary.ResearchResearchCurated research on execution authority and governed autonomy.ContactContactContact routing for Company AI OS review, HELM evaluation, security, and collaboration conversations.

FAQ

Contact FAQ

Can I evaluate HELM without a sales call?

Yes. HELM AI Kernel is public and Apache-2.0. Clone the repository, read the documentation, and run the kernel locally.

What is HELM AI Enterprise access?

HELM AI Enterprise is the reviewed-access company layer around HELM AI Kernel for teams putting AI into governed workflows.

What should I include in my message?

Tell us what AI may propose, what it must never do without approval, and what proof you need.

How fast will I hear back?

Use the category field so requests can be routed correctly. Security disclosures should use the security channel and should not include secrets, customer data, or regulated documents.