Skip to content
PERMISSION/PROTOCOL

Press & media

The facts.
Ready to share.

Company background, product visuals and brand assets for covering Permission Protocol.

Media resources

Download press kit ↓

Full logos, shareable graphics, product screenshots and company copy. No signup required.

Updated

About Permission Protocol

One sentence

Permission Protocol provides external authorization and signed evidence for consequential AI agent actions.

Company description

Permission Protocol helps teams control which consequential actions their AI agents may take and inspect evidence of what was authorized. Its product entry points include GitHub approval workflows through Deploy Gate and MCP tool-call controls through MCP Guard. Teams can define an approval boundary, route requests for a decision and inspect signed authorization receipts. Enforcement depends on the configured integration; a receipt records authorization rather than proof that an action executed. Learn more at permissionprotocol.com.

Brand assets

The full logo pairs the gate symbol with the Permission Protocol wordmark. Use it on first introduction. Keep the proportions and clear space; use the light logo on dark backgrounds and the dark logo on light backgrounds.

Full Permission Protocol logo, light version

Full logo · light

Full Permission Protocol logo, dark version

Full logo · dark

Standalone gate symbol

light gate symbol

Gate symbol · light

dark gate symbol

Gate symbol · dark

Transparent SVG and PNG files. Use these assets to identify Permission Protocol in editorial coverage; do not alter the mark or imply an endorsement.

Ready to share

Make the value visible.

Use these graphics in an article, a presentation or a social post. Download a ready-to-use PNG or the editable SVG.

Your agents take action. You decide.

1200 × 630

Your agents take action. You decide.

Brand overview for articles, presentations and landscape social posts.

From request to decision to evidence

1200 × 630

From request to decision to evidence

Conceptual approval flow. An explanation of the workflow, not a live receipt or execution record.

Before it ships, who approved it?

1080 × 1080

Before it ships, who approved it?

Square graphic introducing the GitHub approval workflow. Enforcement requires the configured approval check and repository protections.

Product visuals

Power to act.
Authority to decide.

Explore a potential Permission Protocol workspace: an action inbox, a deliberate approval and a receipt of the decision. These UI concepts show a product direction with sample data; current functionality is shown in the demo screenshots below.

Product UI concepts · AI generated · Sample data

Potential Permission Protocol action inbox with requests awaiting review, policy-cleared activity and connected systems.

One place to see what needs you.

An authority console brings agent requests, policy decisions and connected systems into one view.

Download full-resolution PNG ↓
Potential approval interface showing a database migration, its exact commit and a named reviewer confirming the change before signing.

Review the exact change. Own the decision.

See the request, understand why it was held and make a deliberate decision tied to its scope.

Download full-resolution PNG ↓
Standalone authorization receipt concept showing Alex Chen approving a sample migration, its repository, commit, policy and decision scope.

A decision you can point to.

A standalone receipt brings the action, approver, policy and exact change together in one record.

Download full-resolution PNG ↓
See actual product screenshots

Captured from the local public demo. Keep the demo designation and identity limitations visible when reproducing these images.

Public demo: an illustrated database migration is presented for review. No production database is connected.

Review the proposed action

Public demo: an illustrated database migration is presented for review. No production database is connected.

Download PNG ↓
Public demo: choose a decision, type a name and confirm before signing. The visitor name is self-reported and unverified.

Make an explicit decision

Public demo: choose a decision, type a name and confirm before signing. The visitor name is self-reported and unverified.

Download PNG ↓
Receipt detail from the local demo, without the surrounding walkthrough. Permission Protocol signs the record. Demo attribution is unverified; this is not evidence of production execution.

The receipt itself

Receipt detail from the local demo, without the surrounding walkthrough. Permission Protocol signs the record. Demo attribution is unverified; this is not evidence of production execution.

Download PNG ↓

A workflow in four frames

From request to record.

Follow one refund through policy, a human decision, an enforcement check and the resulting evidence. Use the sequence in an article or presentation, or download individual frames.

AI-generated product concepts with illustrative data. The payment workflow, identity, verification and adapter results are design examples, not current product screenshots.

Refund policy concept: an $80 refund is allowed, a $4,800 refund needs review, and a new payment destination is blocked by policy.
01 / authority · Product concept
Allow refunds up to $100 to the original payment method. Hold larger refunds for review. Block new destinations, even when a reviewer is available.
AI-generated product concept with illustrative data. Shows a proposed refund policy, not a currently available payment integration.
Approval concept for a $4,800 Acme refund with billing context, original payment method, expiry and Maya Chen identified as the support-lead reviewer.
02 / decision · Product concept
Maya Chen reviews the $4,800 request, its billing context and its original payment destination. Her decision applies to this request; a changed amount or customer needs a new decision.
AI-generated product concept with illustrative data. The reviewer identity and approval state are examples, not evidence of authentication or a real decision.
Enforcement concept comparing an authorized $4,800 refund with a changed $5,200 request that is blocked because the amount no longer matches.
03 / enforcement · Product concept
Approval for $4,800 does not cover $5,200. Check scope and validity before release, through a protected execution path. A new destination remains prohibited by policy.
AI-generated product concept with illustrative data. Release checks and replay protection are depicted design requirements, not verified implementation claims.
Action-history concept separating Maya Chen’s $4,800 authorization from a payment adapter’s report that the provider accepted the refund.
04 / evidence · Product concept
Keep the authorization and the provider-reported result separate. A signed decision records permission; provider acceptance alone does not establish that funds have settled.
AI-generated product concept with illustrative data. Signature verification and payment-adapter results are illustrative; this is not an actual signed receipt or payment record.
Download the full press kit with all four frames ↓

Founder

Rod Carvalho

Founder & CEO · Serial entrepreneur

Rod Carvalho is the founder and CEO of Permission Protocol, a serial entrepreneur and an alumnus of Carnegie Mellon University’s Tepper School of Business. He previously founded BlackLocus, an enterprise machine learning company acquired by The Home Depot. His experience spans enterprise machine learning, distributed systems and marketplace platforms, including distributed ledger work at Flying Manta. At Permission Protocol, he is focused on external authorization for consequential AI agent actions: giving teams control over what agents may do and signed evidence of the decisions behind those actions.

View LinkedIn profile ↗

Primary references

Media contact

Working on a story?

For interviews, product context or additional assets, include your publication, topic and deadline.

[email protected]

Rod Carvalho · Founder