Integration model

Built on the open primitives evidence actually needs.

TimeToPoint does not replace observability, identity, finance or workflow tools. It adds reviewer-readable evidence on top of the open primitives — OpenTelemetry GenAI, MCP, Sigstore, transparency logs, C2PA — with cryptographic signing and transparency-log anchoring shipping as deployment scope expands.

The better question is not "how much can we collect."

The better question is: what evidence does the reviewer actually need?

TimeToPoint starts with one workflow and one review question. From there, it identifies which signals matter and which data should stay out of scope.

Technical primitives — what TimeToPoint adds.

OpenTelemetry GenAI signed wrappers

Reviewer-readable wrappers around OTel spans, not competition with them. OpenTelemetry's GenAI semantic conventions are still in development. TimeToPoint adds the reviewer-readable evidence wrapper the business reviewer needs.

MCP sidecar attestation

Receipts for MCP tool calls. MCP makes tool use easier to standardise. TimeToPoint emits reviewer-readable receipts for MCP tool calls: request hash, result hash, agent identity, timestamp, tool name, authority context.

Sigstore / cosign / in-toto

Cryptographic signing as deployment scope expands. Every evidence record can be signed using Sigstore / cosign with an in-toto Statement v0.1 envelope and anchored to a transparency log.

C2PA Content Credentials

AI-produced media gets a content credential in higher-assurance deployments. Relevant for EU AI Act Article 50 transparency obligations.

DID / Verifiable Credentials

Long-horizon: agent identity recorded as a Decentralised Identifier (DID), with verifiable credentials for delegated authority. Bridge to AmazeWallet.

Existing system categories — what TimeToPoint connects.

Workflow systems

Capture task status, approval states, review points, exception markers, and completion events.

AI agent platforms

Capture authorised tool use, generated outputs, escalation events, human corrections. Native support for LangSmith, Langfuse, AgentOps, W&B Weave.

Observability systems

Reuse selected logs and traces from Datadog LLM Obs, Galileo, Arize, Fiddler. TimeToPoint adds the reviewer-readable wrapper.

Finance and billing

Link invoices, approvals, and payments to evidence of completed, reviewed, and accepted work.

Identity and access

Connect actor, role, permission, delegation, and authority signals.

Document and evidence stores

Retain outputs, approval artefacts, deliverables, and reviewer-facing bundles.

A focused integration path.

1

Select one workflow

2

Identify the reviewer

3

Define the evidence question

4

Map source systems

5

Connect only required signals

6

Generate the record

7

Review usefulness

8

Expand when evidence justifies it

Bring us the workflow. We'll map the evidence path.