> For the complete documentation index, see [llms.txt](https://documentation.rorulabs.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://documentation.rorulabs.com/product/how-it-works.md).

# How It Works

Roru AI routes every AI workflow through a verifiable sequence of cryptographic checkpoints — from session authentication to job completion. Each stage produces proof. No plaintext is accessible outside the enclave at any point in the chain.

## 1. Attested Entry

* **Remote Attestation:** On session start, the local enclave generates a hardware-signed attestation quote proving it is running approved, untampered firmware. The Console does not boot until this check passes.
* **Key Custody:** Signing keys never leave the enclave. Even the orchestrator can only ask for capabilities, not raw secrets.
* **Session Capsules:** Each session spins up a capsule with your policies, access graph, and temporary compute budget.

## 2. Intent Compilation

* **Conversational Entry Points:** Speak naturally. The console compiles statements like “Prep the investor brief and lock down access” into a job graph.
* **Zero-Trust Orchestration:** Jobs are handed to autonomous agents that must present fresh attestations before execution.
* **Deterministic Playbooks:** The compiled workflow is hashed and committed to the audit trail. You can prove what was requested without revealing payloads.

## 3. Secure Data Ingestion

* **Circuit-Level Integrations:** Roru AI threads together cloud apps, on-chain feeds, and air-gapped stores without breaking the perimeter.
* **Policy Enforcement:** Each integration is tagged with purpose-bound policies—intent, scope, retention. Agents that deviate get unplugged.
* **Hermetic Storage:** Retrieved intel lives inside enclave storage encrypted by your keys. Not even cloud providers or host operators get in without you.

## 4. Execution Inside TEEs

* **Agent Runtime:** Autonomous executors operate entirely in TEEs. They receive only the context they need and emit encrypted results.
* **Real-Time Shadow Ops:** Collaborators can co-pilot jobs in a live encrypted workspace. All presence indicators are masked until mutually revealed.
* **Self-Healing Automations:** Workflows monitor themselves, restart failed steps, and keep receipts for post-job audits.

## 5. Signal Extraction & Delivery

* **Encrypted Insight Feeds:** Results stream back as encrypted insights. Your console decrypts and renders them locally.
* **ZK Proof Trails:** Each job produces zero-knowledge attestations showing compliance, integrity, and policy adherence.
* **Selective Disclosure:** Share redacted views with stakeholders, or keep everything locked in the vault.

## 6. Exit Without Footprints

* **Ephemeral Compute:** On job completion, enclave memory is wiped, and the capsule dissolves.
* **Audit on Demand:** Need to prove what happened? You can replay job hashes and attestation proofs without leaking content.
* **Re-Jack Seamlessly:** Saved job templates let you re-run complex operations without re-authorizing every service.

## Key Components

* **Roru Console:** Your operator bridge for jobs, agents, and encrypted coordination.
* **Enclave Integrity Monitors:** Continuous evaluation models that score agents for policy adherence, behavioral consistency, and operational stealth — emitting zk compliance receipts after each stage.
* **Signal Lattice:** A private analytics fabric that turns noise into actionable intel without bleeding data.
* **Executor Swarm:** Autonomous agents that enforce policy using TEE-backed guardrails.

Ready to go deeper? [Get Started](/guides/getting-started.md) or explore the [Feature Reference](/product/features.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://documentation.rorulabs.com/product/how-it-works.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.