Integrations
Push, pull, and the one thing we do not do
Your CRM is not the only system that cares when a deal moves. Indraft tells you when something changes, in whichever direction suits the thing listening, and says plainly where the gaps are.
Push
Webhooks for every event
Register an endpoint and Indraft posts to it when a record is created, updated, archived, merged, or reverted, when an opportunity changes stage, when a task completes, and when an interaction lands. The payload is the change, not a nudge to go and fetch it.
Every delivery is signed. The signature covers the timestamp and the
body together, in the form
t=<unix>,v1=<hmac>, so a captured delivery
cannot be replayed later against a receiver that checks freshness.
Verifying it is five lines in any language and needs no SDK of ours.
A failed delivery retries on an exponential backoff that reaches an hour, across enough attempts to ride out a deploy without hammering an endpoint that is genuinely gone. You can see every attempt, and replay any event, without making the change again.
| Endpoint | What it does |
|---|---|
| GET /v1/webhooks | List registered webhook endpoints |
| POST /v1/webhooks | Register a webhook endpoint, returning its signing secret once |
| DELETE /v1/webhooks/{id} | Disable a webhook endpoint |
| GET /v1/webhooks/deliveries | Recent delivery attempts and their outcomes |
| POST /v1/webhooks/events/{id}/replay | Queue an event for delivery again |
Pull
Or read the change feed and stay in control
/v1/changes is every change in the workspace,
newest first, cursor paginated. Nothing to host, nothing to expose,
and a job that was down for a weekend catches up by asking again from
where it stopped.
The feed is deliberately at-least-once. Ask for changes since a timestamp and you may see one you already have, because the alternative is a change landing in the same instant as your cursor and being skipped forever. A duplicate is a nuisance; a silent gap in a customer record is not.
- Webhooks
- You host an endpoint. Indraft pushes, signs, and retries. Best when something has to happen promptly.
- Change feed
- You poll. Nothing to expose, and catching up after an outage is the same call you were already making.
- Both
- The webhook wakes your job and the feed is what it reads, so a missed delivery costs latency rather than data.
What is not here
There is no importer
No CSV upload, no one-click migration from another CRM, and no prebuilt connectors to a mail provider or a calendar. Indraft has an API and an agent that can call it, and for most people that is the import: point the agent at the spreadsheet or the old system and let it write, which produces a CRM with attribution on every field rather than ten thousand rows nobody can explain.
That is a real difference from the CRM you are leaving, and it is worth knowing before you sign up rather than after.
If you are not leaving it, that is the more common case and it has its own answer: keep your CRM and put an agent beside it. The incumbent stays the system of record, the agent writes here, and the nightly import cannot undo what it learned.
Getting your data back out
One call returns everything: records, history, provenance, and configuration, in a documented shape. It is not a courtesy feature we will remove, and it is reachable both by a person and by a backup job holding a credential scoped to nothing else. Above 100 MB in one file it refuses rather than returning part of your workspace, and we run it for you.