Skip to main content
Glama

Save a concept (agent-authored)

flypost_save_concept
Destructive

Save a concept whose WORDS you wrote yourself: carousel slides, or the title/angle brief for a single image or infographic. Lands at concept stage with no flypost_develop_concept pass (and no AI billing). Returns editor_url — the in-app editor for this concept. Single images: set image_substyle AND the matching image_copy (the exact words baked onto the image) — the render uses those verbatim; without them the image model invents the headline and can feature the wrong number. Infographics: same deal — set infographic_style AND the matching infographic_copy, which carries every figure and line the graphic shows. The prose angle does NOT reach the image model intact, so any number you leave only in the angle gets invented. Then follow the returned next_steps: by default flypost_design_visual designs it (a carousel deck, an infographic, or a stat or comparison Image post) in their brand; the template deck and AI image pipelines are for when the user asks for them. Pass content_bank_item_id to overwrite an existing item in place; omit it to create a new one. Ground the copy via flypost_get_writing_voice + flypost_search_source_docs first. The angle you pass is automatically filed into a per-agent Brain source package so your framing stays visible and checkable — pass agent_name so it is filed under your name.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "title": "<title>",
      "target_format": "<target_format>"
    },
    "name": "flypost_save_concept"
  }
}

Input Schema

NameRequiredDescriptionDefault
angleNoThe angle / narrative context (2-3 sentences).
titleYes
slidesNoCarousel only — the deck you wrote. Order: one hook (numbered-list headline + short subtitle), optional rehook, body slides (auto-numbered), optional recap, one cta last. Keep each slide ~22 words.
sourcesNoSources the copy is grounded on (slide source_refs point at these ids).
signatureNoDeck accent colours; defaults to the brand palette.
agent_nameNoYour product name as the calling agent (e.g. 'Claude'). Names the per-agent Brain source package your material is filed into and the 'Idea by …' attribution shown in the post editor. Defaults to 'AI agent'.
image_copyNosingle_image only — the EXACT words baked onto the image (required for single images). Fill ONLY the fields for the chosen image_substyle: striking_stat → stat (+ label); comparison → before + after (+ header); meme → setup + punchline; product_in_situ → annotation (+ label); ai_decide → picture + words. Keep every line SHORT (one line max; the ai_decide picture may run to two sentences) — it is rendered on the image, not a paragraph. The striking_stat hero line is a complete claim with the number embedded in it (never the bare figure), and its number must be YOUR conclusion, never a figure you are debunking. Without this the image model invents the headline (and can feature the wrong number).
project_idNoProject id; omit for the active project.
target_formatYes
authoring_noteNoOne short sentence on what this is based on (e.g. the conversation or document it came from). Shown as attribution in the post editor.
image_substyleNosingle_image only: striking_stat (one striking stat line IS the whole point) / comparison (two named sides) / product_in_situ (an AI-drawn product scene + callout; it redraws the screen, so for the user's real screenshot use flypost_design_visual with images) / meme (a joke) / ai_decide ("Let AI decide": a picture-led image, for an idea better shown than stated when none of the others fits). Pick the one the idea can actually deliver, then supply the matching image_copy. Defaults to striking_stat.
format_archetypeNoCarousel template archetype; defaults to listicle.
infographic_copyNoinfographic only — the EXACT body copy baked onto the image (required for infographics). Fill ONLY the array for the chosen infographic_style: stats → points; list → items; how_to → steps. Every value is rendered as given, so write finished copy and keep each line short. CRITICAL for stats: the render shows ONLY the figures you put in points[].value — anything you leave in the prose angle never reaches the image model, so a figure that is not here will be INVENTED by it. Pass every number the graphic should show, and no number you cannot source.
infographic_styleNoinfographic only: stats (figures by category) / list (numbered points) / how_to (ordered steps). Defaults to stats. Supply the matching infographic_copy.
source_package_idNoOptional notebook this concept is grounded on.
content_bank_item_idNoUpdate this existing item in place; omit to create a new concept.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.5/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations flag this as a non-read-only, destructive write; the description adds behavior beyond that — no AI billing, returns editor_url, the angle is filed into a per-agent Brain package, and the destructive path is spelled out ('overwrite an existing item in place' via content_bank_item_id). It stops short of saying whether that overwrite is reversible or listing rate/permission constraints, so it is strong but not complete.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the save action, then the per-format safeguards, then next_steps, then the overwrite and grounding notes — a logical order with no filler paragraphs. It runs long for a save tool and restates the 'image model invents the number/headline' warning once for images and again for infographics, which is the only real trim opportunity.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With 16 parameters, nested copy objects and no output schema, the description carries the right load: it names the return value (editor_url), the downstream call (next_steps -> flypost_design_visual), and the failure mode of omitting copy. Only the consequences of overwriting an existing item are left unstated.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is already 88%, so the baseline is 3; the description earns more by supplying cross-parameter coupling the schema only implies — image_substyle must be paired with the matching image_copy fields, infographic_style with the matching infographic_copy array, and the prose angle does not reach the image model. That is genuine coordination guidance, not restatement of field text.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('Save a concept whose WORDS you wrote yourself') and immediately enumerates the three supported shapes (carousel slides, single-image brief, infographic brief). It also draws the boundary against flypost_develop_concept ('lands at concept stage with no flypost_develop_concept pass') and flypost_design_visual, so an agent can pick it without opening a schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicit routing: use this when the copy is agent-authored, follow with flypost_design_visual by default, use template/AI image pipelines only when the user asks, ground via flypost_get_writing_voice + flypost_search_source_docs first, pass content_bank_item_id to overwrite and omit to create. That is when/when-not plus named alternatives.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.