Tesseract

Make a video clip

generate_video

Make a short clip from words — with its own sound and spoken lines (write the dialogue in quotes) — optionally starting from a picture or keeping people and things looking like reference pictures, with Tesseract (charged to the signed-in account). Default gemini-omni-1.2-flash (3–15 s, up to 6 reference pictures). Returns a job_id; video takes minutes — call check_job every 10–20 seconds.

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": {
      "prompt": "<prompt>"
    },
    "name": "generate_video"
  }
}

Input Schema

Name	Required	Description
`model`	No	A video model from list_models.
`params`	No	The model's own fields (fal.ai's names), e.g. {"aspect_ratio": "16:9", "seed": 7}. A field the model's price does not cover is refused, never sent.
`prompt`	Yes	What happens, and how it is filmed.
`max_usd`	No	The most this may cost, in US dollars: a dearer job is not started.
`rights_ok`	No	With a source video of someone else's: true only when the person says the video and its sound belong to their makers, and the result is for their own use unless they hold the rights.
`first_frame`	No	A picture: a public https:// URL, or a data:image/…;base64,… URI (up to 10 MB).
`aspect_ratio`	No	e.g. "16:9", "9:16".
`duration_sec`	No	Its length; each model makes its own set of lengths.
`likeness_consent`	No	Only when the person asked for it themselves, about who is shown: "self" (them), "consented" (an adult who agreed to it) or "made-up" (a made-up person) — and only when nobody shown is under 18 or a public figure, and there is no nudity or sexual content. Needed by the effects that put a person in, talking or lip-synced people, headshots and try-ons.
`reference_images`	No	Gemini Omni: pictures of who or what must look the same in the clip (a character's face, a product) — not frames; up to 6 on gemini-omni-1.2-flash.

Schema Changelog

Changes observed during successful MCP inspections.

Changed2 schema fields changedNov 10, 2026

addedInput schema / properties / likeness_consent

Added value: +{
+  "description": "Only when the person asked for it themselves, about who is shown: \"self\" (them), \"consented\" (an adult who agreed to it) or \"made-up\" (a made-up person) — and only when nobody shown is under 18 or a public figure, and there is no nudity or sexual content. Needed by the effects that put a person in, talking or lip-synced people, headshots and try-ons.",
+  "enum": [
+    "self",
+    "consented",
+    "made-up"
+  ],
+  "type": "string"
+}

addedInput schema / properties / rights_ok

Added value: +{
+  "description": "With a source video of someone else's: true only when the person says the video and its sound belong to their makers, and the result is for their own use unless they hold the rights.",
+  "type": "boolean"
+}

First observedNov 7, 2026

TDQS

A4.6/5.0

Behavior5/5

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

Annotations only declare it as an open-world, non-idempotent, non-destructive write. The description adds substantial context on top: billing to the signed-in account, model default and its 3–15 s / 6-reference limits, that videos take minutes, and that it returns a job_id requiring check_job polling. Consent and rights semantics for people shown are also flagged.

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?

Dense but front-loaded — the core action comes first, then options, then the async follow-up. The heavy em-dash parentheticals make a single long sentence slightly harder to scan, but virtually every clause carries operational information.

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

Completeness5/5

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

For a 10-parameter, nested-params, open-world generation tool with no output schema, the description covers the missing pieces: what comes back (job_id), how long it takes, how to poll, and the cost/consent constraints. Nothing an agent needs to call it correctly is absent.

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 100%, so the baseline is 3. The description still adds real meaning: dialogue goes in quotes inside the prompt, reference_images are likeness references rather than frames, and the duration and reference-count bounds of the default model. It does not explain max_usd or rights_ok beyond the schema, but overall it beats the baseline.

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 ('Make a short clip from words') plus the generative dimensions (sound, spoken lines in quotes, optional first frame, reference-image likeness). 'Short clip' plus the async job model distinguishes it from make_long_video, and the image references make clear it is not generate_image or edit_image.

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

Usage Guidelines4/5

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

Gives concrete usage cues: write dialogue in quotes, start from a picture via first_frame, use reference_images to keep people/things looking the same, and poll with check_job every 10–20 seconds. It never explicitly excludes a sibling by name (e.g. when to pick make_long_video or apply_effect), so it falls short of a full when/when-not statement.

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

All Tesseract tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?