Integrations¶
AnvaiOps connects to the systems where your incident and code knowledge already lives. Every connector is read-only — it observes your source and never writes back.
Adding a connector
Connectors are added and authorized from the console. Each needs an OAuth authorization or API token with read scope on the content you want indexed.
Available connectors¶
| Source | Indexes | Auth |
|---|---|---|
| Jira | Issues, comments, resolutions | OAuth |
| Slack | Channel threads + resolutions | App install |
| Confluence | Spaces, runbooks, postmortems | OAuth |
| Zendesk | Tickets + resolution notes | API token |
| PagerDuty | Incidents, timelines | API token |
| GitHub | Issues, PRs, code (code memory) | GitHub App |
| ServiceNow | Incidents, KB articles | OAuth |
| Datadog | Monitors, incident context | API + app key |
(Connector availability varies by plan. See anvaiops.com/pricing for the per-tier matrix.)
How ingestion works¶
flowchart LR
SRC[Source API] -->|read-only pull| CONN[Connector]
CONN -->|normalize| DOC[AnvaiDocument]
DOC -->|ingest| ENG[ProximaDB engine]
ENG -->|hybrid index| Q[Searchable memory]- Pull — a scheduled job reads new/changed items since the last watermark.
- Normalize — each item becomes a uniform
AnvaiDocument(title, body, source metadata, links). - Ingest — documents are chunked, embedded, and indexed once. Queries reuse the index, so you pay for ingest once.
Data handling¶
- Read-only. No connector has write scope on your source.
- Tenant-isolated. Every document is scoped to your tenant; queries can only reach your own data.
- In your region. Ingested data lands in your cloud's region and stays there.