Integrations / Connect your data and tools

Connected tools.
Traceable results.

Bring search data, AI models and your mailbox into one prospecting process. Every integration needs a clear configuration and source so real results remain distinguishable from demonstrations.

Concept illustration: search, AI and email tools connected through a data hub
Connect deliberately

Know what each integration requires.

An adapter does not mean a service is enabled for your tenant. Use configuration and test results in your workspace as the reference.

Configuration required

Search & maps

Serper / Google Maps / Amap

Discover companies and geographic context using authorized interfaces, credentials and provider quotas.

Provider authorization

Company & social data

Regional search / Crunchbase / LinkedIn

Configurable adapters or demonstration flows may be available. Commercial platforms require the relevant rights; their presence is not a promise of unrestricted collection.

Configuration required

AI models

OpenAI / DeepSeek / compatible

Use OpenAI, DeepSeek or another OpenAI-compatible model for qualification and email drafts. Inspect evidence and source labels. Missing configuration can produce explicitly marked mock or fallback output.

Configuration required

Email delivery

SMTP / Amazon SES

Connect your own sender accounts and apply review and sending policies. Deliverability depends on mailbox configuration and channel rules.

Tenant permissions

Internal systems

API / Webhook

Connect internal processes using scoped credentials. Validate authentication, event types, replay and idempotency to avoid repeated processing.

Planned direction

Orchestration extensions

LangGraph / Harness / MCP Gateway

The present workflow uses an internal state machine and adapters. These ecosystem extensions are not claims of accepted production integrations.

Before operational use

Start with one test record.

Confirm permissions, configuration, source labels and failure handling before adding a provider to live workflows.

  1. Have an administrator configure tenant-level credentials.
  2. Test the connector or mailbox and inspect the response.
  3. Check real / mock labels and failure reasons.
  4. Validate collection, scoring, review and delivery with a small sample.
Questions, answered

A few things to know.

What does source=mock mean?

It marks demonstration or fallback output. It does not mean a third-party provider has returned real business data. After adding credentials, verify authorization, quotas, returned content and failure behavior.

Which AI models can I use?

Choose OpenAI, DeepSeek or any provider that supports the OpenAI Chat Completions contract. Configure its model name, API key and optional base URL.

Are LangGraph and MCP Gateway production integrations?

Current workflows use an internal state machine and adapter interfaces. LangGraph, Harness and MCP Gateway remain expansion directions in the product plan, not claims of accepted production integrations.

Can APIs and webhooks connect internal systems?

Use tenant-scoped credentials and the deployed API contract. Validate event types, authentication, retries and idempotency before connecting operational systems.

Find the fit. Build the relationship.

Make the next conversation count.

Start with a clear market. Connect your research, outreach and team follow-up.

Existing users can sign in. New team access is provisioned by the deployment administrator.