Glossary
The vocabulary Coppica uses across the app, the API, and these docs. If a term in the API reference is unfamiliar, it is probably defined here.
The loop
Coppica’s core cycle: generate copy, track which copy actually converts, update the taxonomy weights from those real outcomes, and let the next generation be informed by what worked. The platform is built around closing this loop rather than generating in a vacuum.
Taxonomy
The catalog of persuasion techniques (1,400+ items) Coppica draws from when generating. Each technique carries weights that shift based on outcomes, so the catalog reflects what converts for a given audience, not generic priors.
Client
The end business you generate copy for. Each client has an intake brief and accumulates strategy outputs, drafts, and tracked outcomes over time.
Intake
The structured brief that describes a client: what they sell, their ideal customer, competitors, and positioning. Intake feeds context into every bot run.
Strategy bot
A run-once bot that produces a durable strategic deliverable (for example a persuasion map, offer, or brand voice). Strategy outputs become reusable context for the writing bots.
Writing bot
A multi-turn bot that drafts copy, grounded in the client’s strategy context. Mimir is the flagship writing strategist; specialist bots cover specific channels and styles.
Output
A single generated piece produced by a bot run. Outputs are the unit that gets tracked, attributed, and learned from.
Draft
An output promoted into an editable, finalizable state. Drafts can be revised and then deployed, at which point they can be tracked.
Outcome
A real conversion event you report back to Coppica (a sale, signup, or other goal). Outcomes are what feed the loop. Without them, generation cannot learn.
Attribution
Linking a conversion to the specific output that drove it. Attribution is what lets the taxonomy update on real performance instead of guesses. Confidence-scaled learning means an outcome’s weight scales with how confident the attribution is.
Tracking link
A UTM-tagged link tied to a specific output, so downstream conversions can be attributed back to the copy that produced them.
Persuasion mode
An orientation toggle (Aggressive, Balanced, or Equity) that shapes which techniques a bot reaches for. Set per client as a default and overridable per session.
Campaign
A container that groups related outputs and tasks for a client, so work and results can be organized around a shared objective.
MCP server
The Model Context Protocol endpoint that exposes Coppica’s primitives to external agents (generating copy, ranking and iterating on outputs, and reporting outcomes) so an agent can run the loop programmatically.
Backfill
Importing copy you already have into Coppica so it can be tracked and learned from, rather than only generating net-new copy.