Every new source is an engineering project
Traditional platforms require a custom connector per source: a scoping call, a statement of work, a deployment cycle. Adding a tool to your stack means waiting on someone else's roadmap.
Wave · AI Source Profiler
Add any content source without an engineer.
Point Wave at a CMS, knowledge base, or video library. Wave's AI reads the vendor's documentation and drafts the ingestion configuration. You preview exactly what would be ingested, then approve. Nothing commits until you say so.
Wave~
Accounts
People
Buying Groups
Schedule
Content
Insights
Settings
Settings › Content sources
New source: help-center (draft)
Source typeKnowledge base
Discoverysitemap
Items found128
Preview ingest
How to configure SSOarticle
Quarterly platform tourvideo
DiscardApprove source
Quick answer
The AI Source Profiler is Wave's no-code path for onboarding a new content source. An operator pastes vendor API documentation or a sample payload; Wave drafts a complete ingestion configuration covering authentication, item discovery, field mapping, and governance; the operator previews the draft against real content and commits it only when it looks right. Nothing auto-commits.
Key capabilities
- AI-drafted source configuration from vendor documentation
- Deterministic preview against real content before commit
- Reference profiles for CMS, knowledge base, and video sources
- Governance classification on every ingested asset
- Automatic monitoring with auto-disable on repeated failures
- Encrypted credentials and allowlisted outbound access
Last updated: July 2026
The problem
Your content lives in five systems. Your intelligence layer sees one.
Blog posts in the CMS, case studies in the help center, demos in the video library, docs in the wiki. Every system your platform cannot read is content your buyers never get recommended.
Invisible content cannot work for you
What your intelligence layer cannot see, it cannot recommend. Unread sources are invisible to gap analysis, buying-group curricula, and per-person predictions, however strong the content.
Stacks change faster than connector catalogs
Teams add new content tools every quarter. Vendor connector catalogs grow by a handful a year. The math never closes.
How Wave does it
From vendor docs to live source in one operator session.
The profiler turns source onboarding into a review task instead of a build task: AI drafts, the operator verifies, the platform runs it.
- 01
Point Wave at the source
Paste the vendor's API documentation or a sample payload into the profiler. That is the entire technical input: no code, no ticket, no scoping call.
- 02
AI drafts the configuration
Wave reads the documentation and proposes a complete source profile: authentication, item discovery, field mapping for title, permalink, format, and summary, and governance classification for every item.
- 03
Preview against real content
A deterministic preview shows exactly which items would be ingested and how every field maps, run against the live source. Nothing is committed at this stage, ever.
- 04
Approve and activate
The operator commits the profile. From then on Wave handles incremental syncs, deduplication across sources, and failure monitoring.
- 05
Content flows into everything else
Ingested assets run through Wave's content tagging automatically, join the content inventory, and become eligible for buying-group recommendations and per-person format predictions.
Where it fits
The front door to Wave's content intelligence.
Content Intelligence can only tag what it can read. The profiler is how everything your team publishes, everywhere it publishes, becomes part of one intelligence layer.
Your content systems
Wave
Tagged, recommendable content
Reference profiles for WordPress, Zendesk, and Wistia ship with the product, so operators see the expected shape before generating their own. Once a source is live, its assets flow through the same tagging pipeline as everything else: themes, personas, funnel stage, and format. Governance classification runs on every ingested item, so drafts, internal documents, and confidential material are structurally excluded from outward-facing recommendations. Credentials are encrypted at rest and never stored inline, and all outbound access runs against a strict allowlist.
Works alongside Content Intelligence, Next Up Content, The full Wave platform.
Why Wave is different
Other platforms send you a statement of work. Wave sends you a draft.
Source onboarding is where content intelligence projects stall. Wave moved it from the engineering backlog to the operator's browser tab.
Most tools
Wave
Each new content source means a custom connector: an engineering sprint, a scoping call, and a deployment cycle.
An operator onboards a source from a browser tab. AI drafts the configuration in minutes; no code is written.
Low-code tools still hand you a mapping spreadsheet and a services engagement.
Wave reads the vendor's own documentation and proposes the mapping. You review a human-readable preview, not a schema.
AI-generated configuration gets applied automatically, and you discover the mistakes in production.
Nothing auto-commits. The preview is deterministic, runs against real content, and an operator approves every profile.
More sources means more silent breakage that nobody owns.
Wave monitors every sync, surfaces failures, and disables a repeatedly failing source instead of quietly ingesting bad data.
FAQ
Questions buyers ask about the AI Source Profiler.
What is the AI Source Profiler in Wave?
It is Wave's no-code path for onboarding a new content source. An operator pastes vendor API documentation or a sample payload, Wave's AI drafts the full ingestion configuration, and the operator previews the result against real content before committing. Nothing is ingested without explicit approval.
Which kinds of content sources can Wave ingest?
Web CMS platforms, knowledge bases, video libraries, and REST-based content APIs. Reference profiles for WordPress, Zendesk, and Wistia ship with the product so operators can see the expected shape before generating their own. If the source exposes documentation or a sample payload, the profiler can draft a configuration for it.
Does the AI ever commit a configuration on its own?
No. The flow is draft, preview, commit, and the middle step is mandatory. The preview is deterministic and runs against the live source, so what you approve is exactly what ingestion will do.
What does the operator review before committing?
A human-readable preview: which items the source would yield, how each field maps (title, permalink, format, summary), and the governance classification each item would receive. If the draft is wrong, edit it or regenerate. Commit only when it is right.
How does Wave handle source credentials?
Credentials are encrypted at rest and never stored inline in the configuration. Wave decrypts them only at use time, and all outbound requests run against a strict allowlist, so a source profile can never route data somewhere it should not go.
What happens when a source starts failing?
Every sync is monitored and logged. A source that fails repeatedly is automatically disabled and surfaced to the operator rather than left to ingest bad data. Re-enable it once the upstream issue is fixed.
What happens to content after it is ingested?
It flows through Wave's Content Intelligence tagging: themes, personas, funnel stage, and format. Tagged assets join the content inventory and become eligible for buying-group recommendations, per-person format prediction, and coverage analysis. Drafts, internal documents, and confidential material are excluded from outward-facing recommendations.
How do I see the profiler on my own sources?
Book a 20-minute walkthrough. Bring the API documentation for a system in your stack and watch Wave draft the ingestion profile live.
See it on your stack
Onboard a content source in your first session.
Book a 20-minute walkthrough. Bring the API docs for a system in your stack and watch Wave draft a working ingestion profile before the call ends.
Request a demo