For film studios, music teams and game publishers
Control what an agent can do before it touches your IP.
An agent can fetch footage, transform a track or publish a game asset in seconds. A log after the fact doesn't stop an unauthorized use. The control point belongs before the action: which asset, which operation, whose approval, and for how long?
The gate
Explicit permission. Limited scope. Fail closed.
Route in-scope agent calls through The Membrane. Each call needs a live, signed, time-bounded scope naming the permitted model, tools and task. Out-of-scope calls are blocked. Expired authorization or broken continuity stops the next call. Each action leaves a tamper-evident, hash-linked receipt.
Licensed content: a worked example
Gate every step. Fail closed on missing provenance.
An agent prepares content from licensed footage, music or game assets. Before it can fetch, transform or publish an asset, operator policy checks the asset's provenance and the permitted operation. If required provenance is missing or unverifiable, no scope is issued and the call is blocked.
Check before granting scope.
Connect your provenance checks to scope issuance. In the proposed Content Credentials workflow, policy checks the manifest signature and signer trust before authorizing the named asset and operation. Provenance is an input to policy, not a substitute for rights clearance.
No proof, no permission.
A missing manifest, invalid signature or untrusted signer fails the policy check. The agent receives no authorization to proceed. A denial is recorded rather than treated as a warning to ignore.
Keep the evidence with the action.
Allowed operations write linked receipts so reviewers can trace an action back to the signed scope that authorized it. A broader operation needs its own authorization; permission to fetch is not permission to publish.
Integration boundary: this is a worked example, not a shipped media connector. The repo ships no C2PA verifier today. Asset provenance checks and media tools must be connected to operator policy; the gate and receipt chain are the existing foundation.
Your assets. Your control.
Authorize before execution. Keep a record you can verify.
An inline control point.
Detection can flag content after publication. The Membrane controls routed agent actions before they execute. It complements detection and provenance systems rather than replacing them.
Self-hosted, operator-owned.
Run the gate in your environment and hold your own signing keys. Define the tools and tasks an agent is allowed to use, with short-lived scopes instead of open-ended permission.
Tamper-evident receipts.
Hash-linked receipts tie actions to their authorization. Export JSON Lines or OCSF-inspired JSON for your review and security workflows, without relying only on a model provider's logs.
Try the authorization layer
Grant. Allow. Block. Sever. Inspect.
The public sandbox shows the gate logic: issue a scope, allow an approved action, block an unauthorized model or tool, and stop an agent when authorization expires or is severed. Inspect the linked receipts and export the evidence.
The sandbox uses simulated Jira, Slack and GitHub tool effects, ephemeral keys and an in-memory bus. It does not process licensed media or hold production credentials. It demonstrates the control layer that a media workflow would use.
Run it locally
No secrets. No paid APIs.
Run the same interactive gate walkthrough on your machine. Gate checks and receipt chaining, with simulated tool effects.
git clone https://github.com/membrane-hq/the-membrane.git
cd the-membrane
cargo run -p membrane-cli --features demo -- demoThen open http://127.0.0.1:8790/