Glossary
One canonical term per concept. Fixes the Workflows-vs-missions vocabulary drift across marketing, docs, and API.
Flowmanner's interface historically said Workflows while the product and API said missions and templates. This is the canonical vocabulary; other words you may meet elsewhere map onto these.
| Canonical term | Meaning | Lives at |
|---|---|---|
| Mission | One executed unit of work moving through a lifecycle (draft → pending → queued → running → completed/approved/failed/paused/aborted/blocked). If marketing copy says "run a Workflow", executing it creates a mission. | app/models/mission_models.py:247 (states at :74) |
| Agent | A configured LLM worker: system prompt, tool access, model parameters. Agents are assigned to missions. | app/models/agent.py:51 |
| Agent template | A reusable, shareable agent configuration. | app/models/agent.py:106 |
| Blueprint | A graph-structured automation definition (nodes/states) executed by the graph engine — the power-user sibling of linear mission templates. | app/models/blueprint_models.py:60 |
| Mission template | A reusable mission definition you instantiate instead of building from scratch. | app/models/mission_advanced_models.py:13 |
| Listing | A template published to the community marketplace, where others can install, rate, and fork it. Ratings and forks follow the trust rules on the Trust Model page. | app/models/models.py:40, app/models/community_models.py:19 |
Terms we deliberately avoid
- "Workflow" — legacy UI/marketing word for what the API calls missions, templates, or blueprints depending on context. Docs use the specific term.
- "Bot" — say agent.
- "Run" as a noun — say mission (or execution when referring to one attempt of a blueprint run).
Model-id spelling
Provider model ids that hit strict upstreams must be fully qualified in
type/model form, e.g. deepseek-chat/deepseek-v4-flash; bare names can be
rejected with a 400. See Models reference for details.
Last updated 2026-08-25 (git-derived)