Skip to content

Integrations

Works with the tools your teams already use.

Claims come from your work trackers. Evidence comes from source control, pipelines, deployments, feature flags, observability and test management. Every tool maps onto one model of work, so the auditor never depends on a vendor’s data shape.

Claims

Work tracking

  • Jira
  • Azure DevOps Boards
  • Linear
  • GitHub Issues
  • GitLab Issues
  • ServiceNow Agile

Evidence

Source control, pipelines, deployments, flags, monitoring and tests

  • GitHub
  • GitLab
  • Bitbucket
  • Jenkins
  • Buildkite
  • Argo CD
  • Kubernetes
  • LaunchDarkly
  • Unleash
  • Datadog
  • Sentry
  • TestRail

One model of work

What the auditor reads

  • Work itemsRevisions, comments, sprints and links
  • Pull requestsWith their reviews
  • CI runsQuality gates and test reports
  • BuildsDigests and provenance
  • DeploymentsEnvironment, version and artifact
  • Flag statesRollout rules and targeting
  • Runtime observationsMetrics, logs, checks and error events
Fig. 1Some of the tools Yakini reads, and the one model each of them maps onto. Every family and tool is listed below.

Claims and evidence

Where Yakini looks.

Each family of tools gives the auditor something specific. Here is what Yakini takes from each, and the tools it reads it from.

  • Claims

    Work tracking

    Where claims come from: status changes, comments, percentages and the acceptance criteria.

    • Jira (Cloud and Data Center)
    • Azure DevOps Boards
    • Linear
    • GitHub Issues
    • GitLab Issues
    • YouTrack
    • ClickUp
    • Asana
    • Trello
    • ServiceNow Agile
  • Evidence

    Source control

    The code at the exact commit that shipped, and the pull requests that brought it in.

    • GitHub
    • GitHub Enterprise Server
    • GitLab
    • Bitbucket (Cloud and Data Center)
    • Azure Repos
  • Evidence

    CI/CD and builds

    Pipeline runs with their quality gates and test reports, and builds with their digests.

    • GitHub Actions
    • GitLab CI
    • Azure Pipelines
    • Bitbucket Pipelines
    • Jenkins
    • CircleCI
    • TeamCity
    • Buildkite
    • OCI container registries
  • Evidence

    Deployment

    What is running in each environment, and which build it came from.

    • GitHub environments
    • GitLab environments
    • Argo CD
    • Flux CD
    • Kubernetes
    • HCP Terraform
    • Terraform Enterprise
  • Evidence

    Feature flags

    Whether a change is switched on, and for whom: targeting rules are checked against the audience a claim is about.

    • LaunchDarkly
    • Unleash
    • OpenFeature (OFREP)
    • Azure App Configuration
    • ConfigCat
    • Split
  • Evidence

    Observability

    Runtime observations, for criteria that need evidence from a running system.

    • Prometheus and Alertmanager
    • Datadog
    • Sentry
    • New Relic
    • Grafana
    • Azure Monitor (Log Analytics)
    • Application Insights
  • Metrics

    Incident management

    Incidents with their service, severity and times, never who was paged.

    • PagerDuty
    • Opsgenie
  • Evidence

    Test management

    Test results from where your teams keep them.

    • JUnit XML uploads
    • TestRail
    • Xray Cloud
    • Zephyr Scale
  • Anything elseThe generic webhookFor in-house and other systems: signed JSON events, verified, stored and audited like everything else.See a signed event
Sample dataEach chip is an example of what a family gives an audit, as the audit cites it.

Connecting

Set up step by step, then kept up to date.

Guided setup asks plain questions and, before anything starts, shows what Yakini will read and what it may write: nothing, unless you turn comments on.

How credentials are kept
  1. Step 01

    Sign in, or use a key

    Sign in with the vendor where it offers that. Otherwise paste a key: it is checked before anything is saved, and the answer says whose it is or what’s wrong, in plain words.

  2. Step 02

    A colleague can finish a step

    No admin rights in that system? A one-time link lets a colleague who has them finish exactly that connection, and shows them who asked, why and what Yakini will read.

  3. Step 03

    Changes arrive as they happen

    Webhooks bring each change as it happens. For Jira, GitHub, Azure DevOps, GitLab, Linear and Bitbucket, Yakini can add its webhook itself when the key has the rights to.

  4. Step 04

    Scheduled syncs catch the rest

    Connections are also synced on a schedule, which catches missed, delayed or out-of-order events and keeps systems without webhooks current.

Sample dataSketches of each step as Yakini shows it.

Anything else

Anything else can send signed events.

The generic webhook, for in-house and other systems. Its events are verified, stored and audited like everything else.

Send work items, pull requests, CI runs, builds, deployments, flag states, runtime checks, sprint updates, release notes and milestone declarations as JSON events signed with HMAC-SHA256.

  • work_item
  • pull_request
  • ci_run
  • build
  • deployment
  • feature_flag
  • runtime
  • sprint_update
  • release_note
  • milestone
  • Each request is signed with the connection’s own secret: X-Yakini-Signature: sha256= and the hex HMAC-SHA256 of the raw body
  • One event per request, or a batch of them
  • Retries reuse the event id, so each event is recorded once
  • An invalid event is refused with what’s wrong, field by field, and nothing is partly accepted
A signed event
POST /api/v1/webhooks/inbound/<connectionId>Content-Type: application/jsonX-Yakini-Signature: sha256=3b9f0c…d27a41X-Yakini-Event-Id: deploy-0042{  "type": "deployment",  "source": "jenkins-prod",  "occurredAt": "2026-11-12T09:41:07Z",  "deployment": {    "externalId": "deploy-0042",    "environment": "production",    "service": "payments-service",    "version": "2.14.0",    "commitSha": "4e1c9a2",    "artifactDigest": "sha256:9b7e…c03a",    "status": "succeeded",    "deployedAt": "2026-11-12T09:41:07Z"  }}
  • Signature checked against the connection’s own secret
  • Recorded once, by its event id
  • Linked to its build by the artifact digest
Sample dataA deployment event, as a deployment tool would send it.

Around the audit

Identity, collaboration and reporting.

Sign-in follows your identity provider, notifications and actions reach people where they already work, and results go out to your own tools.

  • Where people work

    Notifications and actions in chat, the verdict on each Jira issue, self-checks from the editor and checks on pull requests.

    • Slack
    • Microsoft Teams
    • Jira issue panel
    • VS Code extension
    • GitHub pull-request checks
  • Identity

    Single sign-on, with sign-in routed by your verified email domains, and provisioning and deprovisioning from your directory.

    • OIDC single sign-on
    • SAML 2.0 single sign-on
    • SCIM 2.0 provisioning
  • Exports for BI tools

    Exports of metrics, work items, verdicts and commitments, once, daily or weekly, without any person-level data, ready to load on your side.

    • Snowflake
    • BigQuery
    • Power BI
  • Publishing and notifications

    Stakeholder updates and briefings published as pages, and signed outbound webhooks for your own systems.

    • Confluence
    • Notion
    • Email
    • Signed outbound webhooks
Sample dataThe work items and messages shown are examples.

Language models

Your choice of model, task by task.

Each organization registers its own providers and chooses the provider and model for each task, including models it hosts itself. A provider receives source code only if you allow it to; an audit stops rather than send code to a model that may not see it.

Providers you can register

  • Anthropic Claude
  • OpenAI
  • Azure OpenAI
  • Z.ai
  • GitHub Copilot
  • Any OpenAI-compatible API
  • Self-hosted models (llama.cpp)
Sample dataOne organization’s choice of model for each task.

How connectors are built

Written from each provider’s documentation.

Tell us which tools and versions you use, and we’ll go through how Yakini connects to each of them.

Talk to us
  • Published documentation

    Each connector is built from the provider’s published API and webhook documentation.

  • Recorded payloads

    It is tested against recorded payloads and local stand-ins of the service.

  • The provider’s own signatures

    Webhooks are verified with each provider’s own signature scheme.

  • Write-back only when you turn it on

    Comments and issues that Yakini writes back to your tracker are off until you turn them on, and each one waits for someone’s approval.

Using a tool you don’t see here?

Tell us what your teams use. The generic webhook takes events from any system, and we’ll go through the options with you.

  • Hosted, dedicated or self-hosted
  • Your choice of models, self-hosted too
  • Pricing set out in a written proposal