Standard
  • Docs
  • Examples
  • Pricing
  • About

Get started

  • Overview
  • Conversation analytics
  • How Standard works
  • Run your first task
  • Connect
  • Join a Space

Build

  • Authoring · Sales calculation
  • Examples
  • Software release verification
  • Operation profiles
  • Integration reference

Concepts

  • Sources and customer vaults
  • Inputs and permissions
  • Accounts and billing
  • Results and lifecycle
  • Verification checks
  • Follow-on tasks
  • Cleanup

Security

  • Security model
  • Threat model and AI risk mapping
Documentation menu

Get started

  • Overview
  • Conversation analytics
  • How Standard works
  • Run your first task
  • Connect
  • Join a Space

Build

  • Authoring · Sales calculation
  • Examples
  • Software release verification
  • Operation profiles
  • Integration reference

Concepts

  • Sources and customer vaults
  • Inputs and permissions
  • Accounts and billing
  • Results and lifecycle
  • Verification checks
  • Follow-on tasks
  • Cleanup

Security

  • Security model
  • Threat model and AI risk mapping
Standard

Managed confidential execution for approved functions.

© 2026 Standard Runtime, Inc. All rights reserved.

  • Security
  • Docs
  • Examples
  • Pricing
  • FAQ
  • About
  • Contact
  • Terms
  • Privacy
  • Cookies
  • Docs
  • Examples
  • Pricing
  • About
  • Get started

    • Overview
    • Conversation analytics
    • How Standard works
    • Run your first task
    • Connect
    • Join a Space

    Build

    • Authoring · Sales calculation
    • Examples
    • Software release verification
    • Operation profiles
    • Integration reference

    Concepts

    • Sources and customer vaults
    • Inputs and permissions
    • Accounts and billing
    • Results and lifecycle
    • Verification checks
    • Follow-on tasks
    • Cleanup

    Security

    • Security model
    • Threat model and AI risk mapping

    Get started

    • Overview
    • Conversation analytics
    • How Standard works
    • Run your first task
    • Connect
    • Join a Space

    Build

    • Authoring · Sales calculation
    • Examples
    • Software release verification
    • Operation profiles
    • Integration reference

    Concepts

    • Sources and customer vaults
    • Inputs and permissions
    • Accounts and billing
    • Results and lifecycle
    • Verification checks
    • Follow-on tasks
    • Cleanup

    Security

    • Security model
    • Threat model and AI risk mapping

    Overview

    Start with one approved function: what Standard manages, what owners control, and how the receiving integration verifies the permitted result.

    Standard is a managed confidential execution service: approved Wasm workloads start just in time inside an attested trusted execution environment, scoped authority and release rules stay outside the caller, and each permitted result includes Standard Proof.

    Keep planning and orchestration in your existing applications, services and agents. Move selected calculations, validations and business rules into approved functions in Standard. The caller requests the work; the accepted function performs the processing, while owner permissions and release rules govern access and disclosure.

    Four protections

    • Authority outside the agent. Owners control which functions may run, which inputs they may use and which results may be released. Standard enforces those permissions independently of the requesting agent.
    • Restricted Wasm execution. Functions use declared input and output interfaces with enforced resource limits. They receive no general shell, filesystem, subprocess or network interface.
    • Confidential execution and key custody. Functions run inside an attested TEE that protects code and inputs during execution. For supported vault and KMS connections, your non-exportable master encryption keys remain in your store. Authorized confidential components may use temporary data keys and connector credentials for permitted processing.
    • Proof before the next step. The receiving integration verifies Standard Proof against the expected operation before it accepts the result. The workflow then evaluates the outcome and enforces permission for the next action.

    Start with Conversation analytics to see the complete explanation, then Run your first task for the setup sequence. The separate function authoring guide explains the Wasm interface using a sales calculation.

    The single-organization case

    Start with one organization and one approved function: Conversation analytics.

    • The Analytics agent requests category counts for the approved batch-042 selection.
    • The input owner authorizes conversation-categories, that batch and permitted-result release to the Analytics integration.
    • The approved classifier processes the conversations inside the protected runtime.
    • The Analytics integration verifies Standard Proof before making the permitted counts available to the agent.
    • The reporting workflow applies its own publication policy.

    The illustrative counts are Coding 420, Writing 310, Research 190 and Other 80. The requesting agent does not receive the raw conversations or source credentials. Changing the input, function or recipient needs applicable owner authority.

    NextConversation analytics

    On this page

    • Four protections
    • The single-organization case
    • Independent participants in one Space
    • What the runtime provides
    • Choose the execution boundary
    • Execution to next action
    • Key terms
    • Continue building

    The example is illustrative. Run your first task explains the setup sequence, including the current hosted continuation limit. Software release verification is a separate example.

    Standard Proof binds recorded execution claims to the permitted result. The receiving integration verifies those claims against accepted references before use, then applies its own next-action policy.

    Independent participants in one Space

    A Space binds the participants, the agreed operation profile, input permissions, required approvals, permitted recipients, and verification expectations for an operation. It can serve one organization; add independent owners only when the function needs their assets or approval.

    • Each participant joins under its own identity and keeps control of its own assets.
    • Each required owner authorizes the use of its own contribution and the release of results. One owner’s approval cannot replace another’s.
    • Sponsoring, inviting or coordinating a Space does not create authority over another owner’s data or code.
    • A recipient receives only its permitted result and verifies Standard Proof against references it accepted itself.

    Example: a customer's maintenance agent asks a vendor to diagnose an issue. The customer contributes an encrypted telemetry window and the vendor contributes its approved diagnostic workload. Each owner authorizes its own contribution and release, and an assurance team verifies the findings with Standard Proof.

    What the runtime provides

    CapabilityPurpose
    Scoped authorityIdentify the caller and establish what it may request, use, or release.
    Policy enforcementApply the accepted rules during execution and before a result leaves.
    Attested runtimeExecute supported workloads within an environment whose identity can be checked.
    Restricted Wasm executionRun the approved function with declared input and output interfaces and resource limits, without a general shell, filesystem, subprocess or network interface.
    Confidential executionDecrypt protected inputs inside the runtime and release only permitted outputs.
    Standard ProofBind the relevant execution claims to a result so the receiving integration verifies the handoff before use.
    Managed lifecycleOperate workload activation, execution limits, result release, evidence, and cleanup, with completion and cleanup reported separately.

    Choose the execution boundary

    Standard runs the approved Wasm workload over authorized inputs inside the confidential runtime. Code, data, and model inputs must fit the workload's supported formats and execution limits. Local encrypted files and supported connected sources supply the inputs. Attested confidential ingress retrieves and validates connected-source material under scoped permission, then contributes it to the operation worker. Customers retain persistent key and API-secret custody in their own vaults or key stores. See Sources and customer vaults.

    Existing external tools remain external. A supported interaction can record what Standard controlled or observed, such as an authorized request and response. That evidence does not attest the provider's internal computation. A tool provider does not need to move its backend into Standard simply because an agent uses its API or MCP interface.

    Select a supported operation whose evidence covers the claim your next step needs. See the integration reference.

    Execution to next action

    1. Configure authority. Define the participants, operation, resources, release rules, and accepted verification references. Each required owner authorizes its own part.
    2. Submit work. Provide the supported operation and input references. Protected contents stay out of ordinary tool arguments.
    3. Enforce the rules. Standard checks authority and applies the execution and release policy.
    4. Return result and proof. The permitted output is bound to execution evidence.
    5. Verify and continue. The receiving integration verifies Standard Proof, then applies its own acceptance policy to the result and enforces permission for the next action.

    Every operation is blocked until each required owner approves it, and an approval covers that operation only, unless that owner has opted in to a standing permission for that exact task (same inputs, recipients and release rule; no connected sources; at most 100 runs over at most 30 days; revocable).

    Key terms

    TermMeaning
    Operation or taskA defined unit of work requested through a supported Standard interface.
    ParticipantAn organization, agent, human or service that contributes inputs or authority, or receives a result. Each keeps its own identity.
    SpaceA governed context for participants, authority, an accepted profile, release rules and evidence. It can serve one organization or several independent owners. It is not a shared running computer.
    Operation profileA versioned, accepted contract for a class of work: its workload, roles, inputs, limits, authority, and result and release rules. It grants no access by itself.
    Per-operation approvalAn owner’s signed approval of one operation request. It does not carry over to the next request.
    Operation requestThe concrete profile reference, workload and input references, parameters, and requested result for one proposed operation.
    Encrypted-input referenceA reference to an encrypted object and its expected digest. Access to the reference does not grant decryption rights.
    Standard ProofExecution evidence binding the relevant runtime, workload, authority, inputs, and result claims within the recorded execution boundary.

    Continue building

    • Run your first task
    • Inputs and permissions
    • Results and lifecycle
    • Security model