Twine operates across Google, LinkedIn, Reddit, HubSpot, and Salesforce without pretending every platform exposes the same capabilities or the same quality of evidence. See what Twine can read, propose, apply, launch, and measure before granting access.
Campaign, creative, targeting, asset, performance, and lead-form evidence through the same Decision Brief and approval model.
Campaign, ad group, post, audience, pixel, funding, targeting, forecast, and performance evidence with platform-specific limits.
Use CRM company, lifecycle, deal, opportunity, and stage data to improve business-outcome measurement and, where enabled, define reviewable company audiences.
Bring the daily read, urgent decisions, account questions, and promised follow-through into the channels where the team already works.
Twine can answer with the surrounding thread intact, remember durable account context, and stage reviewable work without turning Slack into an unreviewed path to spend.
Twine labels evidence as a cached database read, a live platform response, inferred, or unavailable. It also shows freshness, fallback behavior, warnings, and limitations when those details affect a decision.
A number without its source and age is an opinion. Twine does not serve opinions as facts.
If your team works in Claude Code, Codex, or plain terminal workflows, Twine meets you there. A CLI and a local MCP server, in controlled beta, let your own agents inspect accounts, pull campaign context, draft work, and stage it for review.
The same rules apply as everywhere else: external tools connect through tenant-scoped, revocable, scope-limited keys. They never touch the ad platforms directly, and approval and application stay in Twine. See the agentic GTM use case.
Twine starts read-only, holds every consequential change to an approval, and keeps an audit record of who did what and when. The full control model lives on Execute.
Capabilities differ by platform, and Twine says so. The account review shows exactly what applies to yours.