Book a demo

Internal tool integrations

Forty tools.
One graph.

Connect your whole stack once.

01 · Why a graph

Why a graph beats a pipeline

Move a row, land a table, export a CSV. None of them let you ask why.

Zapier moves a row from A to B, Fivetran lands raw tables in a warehouse your data team still has to model, and a CSV export is a snapshot that’s stale the moment it downloads. But Vernais connectors don’t just move data, they resolve it: a GitHub commit, a Jira issue, and a Stripe charge that touch the same release version become one investigation, because they resolve to the same Entity in one graph. No pipeline to maintain, no warehouse to model, no analyst in the loop.

02 · The catalog

Forty connectors, grouped the way you think.

Each one brings its records into the same graph, and its shared things (a customer, a version, a service) resolve to the same Entity across every other tool.

22 live at launch18 on the way40 in one graph

01

Project management

Brings to the graph The work: issues, sprints, and the release versions that tie planning to outcomes.

At launch 2
  • Jiraissues · statuses · versions · work items
  • Linearissues · cycles
Coming later 2
  • AsanaSoon
  • ClickUpSoon
02

Analytics & data

Brings to the graph Behavior and the warehouse: the events, funnels, and modeled tables that show what users actually did.

At launch 5
  • Amplitudeevents · users · funnels
  • Mixpanelevents · cohorts · conversions
  • Segmentevents · identities · traits
  • Snowflaketables · query results
  • BigQuerydatasets · tables · query results
Coming later 6
  • Google Analytics 4Soon
  • HeapSoon
  • PostHogSoon
  • DatabricksSoon
  • Amazon RedshiftSoon
  • PostgreSQLSoon
03

CRM & support

Brings to the graph The customer’s voice: deals, tickets, conversations, and the accounts behind them.

At launch 4
  • Salesforceaccounts · opportunities · contacts
  • HubSpotdeals · companies · contacts
  • Zendesktickets · ticket metrics · requesters
  • Intercomconversations · contacts
Coming later 3
  • PipedriveSoon
  • Zoho CRMSoon
  • FreshdeskSoon
04

Finance & billing

Brings to the graph The money: charges, invoices, subscriptions, and the revenue side of every event.

At launch 1
  • Stripecharges · statuses · customers
Coming later 4
  • NetSuiteSoon
  • QuickBooksSoon
  • ChargebeeSoon
  • PaddleSoon
05

Engineering

Brings to the graph What shipped and what broke: commits, deploys, errors, and the infrastructure signals behind them.

At launch 5
  • GitHubcommits · pull requests · issues · releases
  • Datadogmonitors · metrics · alerts
  • Sentryissues · error events · releases · affected users
  • AWS CloudWatchmetrics · alarms · log events
  • GCP Cloud Monitoringmetrics · alert policies · uptime checks
Coming later 2
  • GitLabSoon
  • VercelSoon
06

Docs & design

Brings to the graph The intent and the discussion: specs, wiki pages, design files, shared docs, and the team conversations around the work.

At launch 5
  • Notionpages · databases
  • Confluencepages · spaces
  • Google Drivefiles · metadata
  • Figmafiles · comments · last-modified activity
  • Slackmessages · channels · threads
Coming later 1
  • Microsoft TeamsSoon

03 · The lifecycle

The connector lifecycle

Connecting a tool isn’t a checkbox. It’s a real, inspectable lifecycle: you choose exactly what comes in, prove it works before you sync, and can walk it all the way back out.

01

Connect

Enter credentials in a form built from the tool itself. Every field maps to what that specific connector needs: a personal access token here, an API key and subdomain there. Secrets are never echoed back; the detail view shows a masked state, nothing more.

  • Data-driven, per tool. No generic “paste your JSON”: the form is shaped by the connector’s real auth requirements.
  • Encrypted at rest. Credentials are stored encrypted, never returned in plain text.
02

Test

Before you sync a single record, Vernais makes one real authenticated call to the tool’s API. You know the connection works because it worked, not because a form validated.

03

Choose fields

Every stream opens into a recursive field tree. Expand an object to see its subfields. Check exactly what you want: a tri-state checkbox rolls selection up and down the tree.

  • Fill-rate bars. They show how often each field actually carries a value in your real data, so you don’t sync empty columns.
  • ID, cursor, and doc badges. They flag the primary key, the sync cursor, and documented-but-unseen fields at a glance.
  • Nested and documented fields. They appear in place, including fields the schema documents but your data hasn’t produced yet.
04

Sync

Vernais pulls live records and projects them down to only the fields you selected (plus the primary key and cursor, always kept). Nothing you didn’t ask for enters the graph.

figma · comments
synced 218 records  ·  source: live
fields kept id, message, created_at
always kept + pk, cursor
05

Disconnect

Remove the connection and its stream selections. Your synced data stays exactly where it is.

  • Cursors reset. The connection is gone.
  • Your graph stays. Everything you already built is untouched.
06

Purge

The full-forget. Deletes the tool’s data across both databases, raw payloads, clean signals, its graph nodes and edges, while preserving shared entities and cross-integration structure.

  • Total for one tool. Purge removes all of that tool’s own data.
  • Cross-tool structure survives. The entities other tools also touch stay resolved.

04 · One shape

One shape for everything

The catalog is forty tools. The graph is one language.

Whatever the source, its records land as Signals: one clean, deduped record each, the atomic unit of evidence. The things those records share resolve to Entities that appear once. ML then groups Signals into Topics with an umbrella→subtopic hierarchy.

Signals

One clean, deduped record each: the atomic unit of evidence.

Entities

The shared things (a customer, a version, a service, an invite code) that appear once, no matter how many tools mention them.

Topics

Signals grouped by ML into an umbrella→subtopic hierarchy.

That’s why a single investigation can span your whole stack. Three tools touch the same release, and land as one story.

Source A GitHub commit on v2.2.1
BecomesA Signal + a v2.2.1 version Entity
Lands inThe same graph as everything else on that release
✓ Built
Source A Jira issue referencing the same version
BecomesA Signal resolved to the shared v2.2.1 Entity
Lands inOne hop from the commit
✓ Built
Source A Stripe charge that failed after the deploy
BecomesA Signal with its outcome and customer Entity
Lands inThe same investigation, timestamped near the others
✓ Built

One commit, one issue, one charge. Three tools, one story, one graph.

05 · Shared, not siloed

Connect a tool once. It’s connected everywhere.

An integration you set up in one workspace shows connected across every workspace you belong to, because a connection is an account-level fact, not a per-project chore. Your data, though, stays strictly isolated: each workspace has its own pair of databases, and membership is re-checked on every call.

  • Connect once. No re-authenticating the same tool in every workspace.
  • Isolated by default. One workspace can never read another’s synced data.
  • Encrypted where it matters. Credentials and chat are encrypted at rest.