Skip to content
FLORA DocsGo to app

Tools reference

Every tool the FLORA MCP server exposes, grouped by resource.

FLORA MCP mirrors the FLORA REST API 1:1 — every API operation is exposed as a callable tool. The agent driving your client decides which tool to call based on your prompt and the tool descriptions, then renders results inline.

Discovering Techniques and starting Technique runs is the core FLORA MCP workflow.

ToolWhen the agent uses it
list_techniques”What Techniques do I have?”, “Find Techniques that do X”
retrieve_techniqueBefore running, to inspect inputs and cost
create_technique_runWhen the user says “run”, “generate”, “make this”
retrieve_technique_runTo poll a previously created run

API Reference: Techniques

Two top-level run endpoints exist in addition to the slug-based Technique runs above. The agent picks them when the user gives a Technique ID rather than a slug, or asks for a generic generation.

ToolWhen the agent uses it
start_technique_runUser refers to a Technique by ID, or chains from earlier results
start_generation_runUser asks for a plain image/video generation without a specific Technique

API Reference: Runs

For workflows that need user-supplied images or videos, the agent creates an asset, uploads bytes via a signed URL, marks it complete, and passes the final URL as a Technique input.

ToolWhen the agent uses it
create_assetUser says “use this image” with a file path; client uploads on user’s behalf
complete_assetAfter upload finishes
retry_assetIf the signed upload URL expired
list_assets”Show me my recent uploads”
retrieve_assetLook up a specific asset by ID

API Reference: Assets

Projects organize Technique runs and uploaded assets inside a workspace. When the agent creates runs, it can optionally tie them to a project so they appear together in the FLORA canvas.

ToolWhen the agent uses it
list_projects”What projects do I have?”
create_project”Make a new project for this campaign”
retrieve_projectInspect a specific project
list_project_nodes”What’s in this project?”
attach_project_asset”Add this output to my campaign project”

API Reference: Projects

ToolWhen the agent uses it
list_workspacesFirst call after OAuth, or when the user has access to multiple workspaces

API Reference: Workspaces

For generic generation workflows that take a model parameter, the agent lists models to discover available IDs.

ToolWhen the agent uses it
list_models”What image models can I use?”

API Reference: Models

ToolWhen the agent uses it
record_feedbackUser says “this one is great” or “this one is bad” about a specific output

API Reference: Feedback

You don’t call tools directly. You describe what you want, and the agent picks a sequence. For example, “Run the Thumbnail Technique with this brief” typically resolves to:

  1. list_techniques (with filter name~"thumbnail") — find the right Technique.
  2. retrieve_technique — see what inputs it needs.
  3. create_technique_run — submit the run with the inputs you provided.
  4. retrieve_technique_run (polled) — wait until it’s complete and render outputs inline.

Different prompts produce different sequences. The Recipes section walks through several end-to-end.

Most clients namespace tools by server. In Claude Code’s /mcp listing they’ll appear as flora:list_techniques, in Cursor as flora.list_techniques, etc. You almost never need to type these — say what you want and the agent picks the tool.

  • Calls that create runs (create_technique_run, start_generation_run) charge your workspace credits. The cost is shown in retrieve_technique under runCost and on the completed run as chargedCost.
  • Rate limits apply at the workspace level and match the REST API. See the API Reference for current limits.
  • A 402 insufficient_credits from a tool means the workspace is out of credits — top up in FLORA → Settings → Billing.