For IT & platform teams

Plug into your stack. Keep the privacy layer.

Null wires into the systems your team already uses, with sensitive data tokenization preserved at every hop, across every source.

Browse catalog Request an integration

Live · Salesforce lead → GPT-5.1 → back to leadFlowing

SalesforceSource

∅

NullTokenize

GPT-5.1Model

∅

NullResponse

ClientReveal

Salesforce · Lead record fetched

Lead Hermann Höfer at Hanseatic Versicherung AG · amount €18,420

Null · 3 entities tokenized · 41 ms

Lead PERSON_A·S7FT9CD at ORG_A·44LT3R · amount AMOUNT_A·18420

GPT-5.1 · Inference · 342 tokens · eu-fra-1

Summarize lead PERSON_A·S7FT9CD, high priority; renewal Q2; AMOUNT_A·18420 at risk.

Null · Response passes through · 0 reveals server-side

Response holds PERSON_A·S7FT9CD · reveal scope declared: thread:sf-lead-8142

Client · Revealed in-browser · audit AE-883104

Lead Hermann Höfer is high priority; renewal Q2; €18,420 at risk.

Round-trip · 412 ms · eu-fra-10 raw sensitive data egress · vault stable

01 · Inline, not sidecar

Tokenization runs in-transit.

No nightly sync. No agent on every laptop. Null sits between the connector and the model, every record is tokenized on its way through, never at rest in a second place.

02 · Auth-aware

OAuth 2.1, SAML, SCIM, handled.

Connectors inherit your workspace's identity provider. Access scopes, role claims and delegated consent flow through to the vault's reveal policy, automatically.

03 · Token stability

One identity. Every source.

The same person gets the same token across Slack, Salesforce and Outlook, so references resolve, audit stays coherent, and cross-system insights keep working.

Token stability

Same vault. Every source.

When PERSON_A·S7FT9CDshows up in a Slack thread, a Salesforce lead and an Outlook subject line, it's always the same token. Your audit log is queryable. Your vectors are clusterable. Cross-source reasoning keeps working, without a single raw name leaving the perimeter.

One vault, every connector

The token vault is shared across integrations. No per-system duplicates, no aliasing drift.

Deterministic, not random

Same raw value → same token, across sessions and workspaces. Embeddings, joins and dedup still work.

Revoke once, gone everywhere

Revoke a token's reveal scope and every connector respects it, immediately, not eventually.

Vault · single source of truth3 systems synced

Slack · #renewals

14:22

Hey Pablo, can you check PERSON_A·S7FT9CD's renewal status? Thread is getting noisy.

Salesforce · Lead

SF-8142

Lead: PERSON_A·S7FT9CD · Company: ORG_A·44LT3R · Amount: €18,420 · Owner: anna.mueller

Outlook · Subject

Inbox

RE: Quarterly review with PERSON_A·S7FT9CD, agenda attached, 2 pages.

Same vault ID · resolves across all three

Catalog

24 connectors today. New ones every two weeks.

Curated, auth-tested, privacy-reviewed. Each connector ships with its entity schema and a dry-run mode so your DPO sees detections before you commit.

4 entity types

Slack

4 entity types

Microsoft Teams

6 entity types

Google Workspace

5 entity types

Notion

7 entity types

Outlook

8 entity types

Salesforce

7 entity types

HubSpot

9 entity types

Dynamics 365

6 entity types

Pipedrive

9 entity types

SharePoint

8 entity types

Google Drive

7 entity types

Dropbox

5 entity types

Confluence

6 entity types

Jira

8 entity types

Zendesk

10 entity types

ServiceNow

7 entity types

Intercom

12 entity types

Snowflake

12 entity types

Databricks

14 entity types

Amazon S3

12 entity types

PostgreSQL

3 entity types

GitHub

3 entity types

GitLab

3 entity types

Linear

Governance

Your SIEM gets the audit. Never the sensitive data.

Every connector invocation, every token reveal, every policy decision streams to your existing log pipeline, with tokens in place of names. Splunk, Datadog, Elastic, or self-hosted, pick one; the event shape is the same.

OpenTelemetry-native

Events emit as structured OTel logs with stable attribute schemas. Your dashboards survive schema changes.

Tokens all the way through

The log pipeline never sees a raw name. Even if your SIEM leaks, your vault holds.

Replay with scope

Export a time range of audit events and replay them against a sealed copy of the vault, for an investigation, for a regulator, for a drill.

POST /v1/audit/event · → Splunk HECStreaming

// one event per connector invocation
event: connector.invoke
ts: 2026-04-21T19:42:11Z
actor: "anna.mueller@aifano.com"
connector: "salesforce.v2"
scope: "lead.read"
subject: "PERSON_A·S7FT9CD"
reveal: "thread:sf-lead-8142"
model: "openai/gpt-5.1"
tokens_used: 342

// never logged: raw name, raw amount

Roadmap

Most-requested integrations. Vote for what's next.

We ship the top-voted connector every two weeks, with the same entity-schema review and dry-run as everything else in the catalog.

Microsoft Dynamics 365

building

Amazon S3

building

Notion

planned

Linear

planned

Zendesk

planned

HubSpot

backlog

ServiceNow

backlog

Auth handshakes

Every connector, pre-signed.

Our security team signs the DPA, reviews the scope, and runs a red-team sweep before a connector ships. Your legal inherits the clearance, no per-integration review cycle.

OAuth 2.1

With PKCE

Every connector

SAML 2.0

IdP-initiated

Workspace SSO

SCIM 2.0

Provision + deprovision

Role claims forwarded

mTLS

Optional

For on-prem hops

Your tool isn't listed? We'll wire it up in two weeks.

Tell us what you need. Null's connector team ships a new integration every sprint, with the full privacy layer, not just a wrapper.

Request an integration Talk to the connector team