Core pillars
Social publishing and scheduling connections
Analytics and attribution ingestion pipelines
Billing, CRM, and notification integrations
Product
Integrations connect each operating station into one system.
Connect social publishing, analytics, CRM, billing, and communication tools so execution can run from one control surface.
We disclose connector categories and data boundaries while keeping failover strategy, secrets handling detail, and anti-replay internals private.
What you get
Supported connector and provider matrix
Data ingress and governance reference
Integration reliability and failure behavior notes
Launch checklist
Validate every workflow claim against live behavior.
Confirm plan limits and metering match pricing page copy.
Publish failure behavior and recovery expectations.
Ensure all operator routes have empty-state guidance.
Link each product page to demo or in-app proof.
Verify primary CTA route: /app
Detail lane
Media realism connector lanes
Potsticker routes creator media through interchangeable provider lanes so teams can balance quality, speed, and gross margin without rewriting the product surface.
Digital twin / talking avatar lanes (for example HeyGen, D-ID style workflows)
Recast and body-swap lanes with motion carryover (for example Higgsfield, Viggle, Reface style workflows)
Reference-to-video lanes for guided scene generation (for example Vidu, Veo, Kling, Seedance style workflows)
Detail lane
Commercial control layer
Live submission is env-gated, broker-routed, and margin-guarded so expensive media lanes do not silently burn profit.
Model broker selects lane by goal: quality-first, balanced, or cost-first
Missing env or planning-only routes are escalated before paid media spend
Attribution and winner scoring feed back into next-batch routing
Detail lane
Open kitchen boundary
We expose connector categories, setup expectations, and operational limits while keeping private ranking weights, anti-abuse logic, and internal optimization recipes confidential.
Public docs describe capability boundaries and failure modes
Private internals include prompt weighting, route scoring, and anti-replay signatures
Operator controls remain visible without leaking private orchestration logic
Common questions
Do integrations run automatically after connection?
Only configured workflows run. Connections unlock lanes, but operators still control publishing, routing, and automation scope.
How are failed webhooks or sync jobs handled?
Events should be idempotent, replay-safe, and visible in operations logs with clear retry and recovery status.
What should we validate before enabling live connectors?
Credential scope, webhook signatures, rate limits, failure alerts, and billing/attribution event mapping.
Do you support recast/body-swap/talking-portrait workflows?
Yes. Potsticker supports those as routed media lanes when the corresponding connector envs are configured and policy checks pass.