Rendley docs

What the agent can do

The automation connector uses the same Rendley Agent as the API. Describe the result in plain language and it works out which editing, analysis, and generation steps are needed.

You do not need to learn a separate command for every operation. The sections below show what the agent supports and which details are worth adding to a prompt.

Understand your source material

Before the agent edits, it can inspect what you gave it instead of treating every file as an anonymous clip.

  • Transcribe audio or video with word-level timing.
  • Find strong quotes, key moments, topics, pauses, filler words, retakes, and unusable takes.
  • Analyze the visual content of images and video, including scenes, objects, framing, shot types, and on-screen text.
  • Analyze music and audio for mood, structure, energy, and useful time ranges.
  • Read a webpage you provide and use its text and available media as source material.
  • Search for relevant stock images, video, and GIFs when you have not supplied enough visual material.

You can ask it to “find the three strongest moments” or “build a product video from this article” without supplying timecodes or rewriting the source.

Edit and compose a timeline

The agent can create a new project or continue an existing one, then perform ordinary timeline work across video, image, audio, text, and subtitle layers.

  • Trim, split, move, delete, and reorder clips.
  • Assemble several files into one story and cut it to a target duration.
  • Remove filler words, long pauses, retakes, and weak sections from spoken footage.
  • Build jump cuts, cutaways, J-cuts, L-cuts, picture-in-picture, split-screen, and grid layouts.
  • Crop, scale, position, rotate, and change opacity.
  • Reframe for vertical, square, or landscape output while keeping important subjects and text in view.
  • Change playback speed and animate clip properties with keyframes.
  • Add text, titles, lower thirds, overlays, logos, shapes, and calls to action.
  • Add transitions, fades, filters, color correction, and color grading.
  • Mix voice, source audio, music, and sound effects; adjust levels and fades.
  • Add styled subtitles or karaoke captions from a transcript.

Ask for the result you care about. “Make this a fast 30-second Reel with the speaker centered and captions inside the safe area” gives the agent more useful direction than a list of low-level operations.

Use built-in production knowledge

Rendley Agent includes production guidance as well as editing tools. It can choose an appropriate structure, pacing, layout, typography, caption treatment, and delivery format for the job instead of waiting for you to specify every creative detail.

Its built-in workflows cover social ads, UGC-style videos, tutorials, testimonials, product showcases, brand videos, explainers, transcript cleanup, subtitles, transitions, color work, animation, and platform-specific output. For complex requests, it inspects the source material first and plans the composition before making the timeline.

You remain in control: provide detailed art direction when you have it, or describe the audience and goal and let the agent propose the execution.

Generate missing media

Rendley’s generation tools feed directly into the editor. The agent can create an asset and place it on the timeline as part of the same job.

  • Generate new images or edit an existing image from a prompt.
  • Generate video from text, a starting image, ending image, reference images, reference video, or reference audio when the selected model supports it.
  • Create voiceover from a script with a selected voice.
  • Generate background music to match a duration and musical direction.
  • Generate sound effects such as impacts, whooshes, ambience, risers, and interface sounds.
  • Source free stock images, videos, and GIFs for B-roll.
  • Build short product spots, social ads, explainers, tutorials, testimonials, carousels, and UGC-style videos from a brief.

Rendley chooses an appropriate model by default. When quality, speed, cost, identity consistency, native audio, or a particular model matters, say so in the prompt. The AI model catalog shows the models currently available through the same backend.

Create motion graphics with your connected model

For animated titles, logo reveals, lower thirds, callouts, charts, transitions, and full motion-design scenes, your connected AI model can author the animation itself.

Rendley gives the model a creative brief, your brand context, and the same motion-design guidance used by Rendley’s own generator. The model writes the editable motion-clip code, Rendley validates it, and the agent places it on the timeline.

This path uses the model access included with your Claude, ChatGPT, Codex, or other client plan and does not consume Rendley generation credits. Normal limits from your model provider still apply. See Motion graphics for details.

Transform and enhance media

The agent can also run AI transformations on existing files and compose the result back into the project.

  • Remove an image background for product cutouts and overlays.
  • Remove a video background to create a transparent subject without a green screen.
  • Improve speech clarity and reduce unwanted background sound.
  • Change a speaker’s voice while keeping the original timing and delivery.
  • Translate and dub a video while preserving the character of the original speaker.
  • Lip-sync a visible speaker to replacement audio.
  • Upscale images by 2× or 4×.
  • Upscale video to a higher resolution, including 4K when the source and selected model support it.

Some transformations have source-duration or model-specific limits. The agent checks the media and tells you when it needs a different approach.

Create captions, versions, and localizations

One source edit can produce a set of versions for different platforms, formats, or markets.

  • Burn in captions using your brand style or a style described in the prompt.
  • Produce clean, readable subtitle tracks from one or several clips.
  • Translate, dub, and subtitle versions for multiple languages.
  • Create 9:16, 1:1, and 16:9 variants from the same project.
  • Make platform-specific cuts with different durations, hooks, layouts, and calls to action.
  • Duplicate a project before making variants so the approved original stays unchanged.

Name the deliverables explicitly. For example: “Create English, Spanish, and German versions in 9:16, with dubbed speech and burned-in subtitles, then export each one.”

Apply your brand kit

A brand kit belongs to a workspace, so you can define it once and reuse it across every project in that workspace.

The agent can read and apply:

  • Brand voice and writing guidance.
  • Named colors and their roles.
  • Uploaded fonts and text styles.
  • Caption styling.
  • Preferred voices.
  • Logos, product images, and other reusable media.

Ask the agent to “use the workspace brand kit,” or call out the exact parts you want it to apply. It can also add colors and upload supported brand assets for future work.

The connector can use:

  • Files on your computer when your MCP client can read them.
  • Files attached to the conversation when the client exposes them to the assistant.
  • Public HTTPS download links.
  • Media already stored in a Rendley project.
  • Logos, fonts, and other assets stored in a workspace brand kit.
  • Text and usable media from a webpage you provide.

You can upload as many as 50 files in one transfer batch. An individual edit can receive up to 20 new files, so larger collections are added in batches and then edited from the project library. See Work with your files.

Manage projects and continue work

The agent can list workspaces and projects, inspect a project, create or duplicate projects, resume the latest conversation attached to a project, and continue refining an earlier edit.

That means the workflow can be iterative:

Open the Launch video project. Keep the current structure, replace the opening with a stronger hook, make the captions larger, and export a square version too.

You can also open the project in Rendley Studio at any time after the agent finishes and continue editing visually.

Export the finished result

The automation connector can render the project and return a temporary download link.

  • MP4 with H.264, or WebM with VP8.
  • 360p, 480p, 720p, 1080p, 2K, or 4K.
  • High, medium, or low quality.

The default is a high-quality 1080p MP4. Download links expire after a few hours, so save the file rather than the URL.

Availability and approvals

Browsing existing projects and reading basic account information are available after connection. Editing, AI processing, workspace management, and export depend on your Rendley plan; generative actions and rendering can consume credits.

Every MCP edit runs interactively. Free timeline operations can continue automatically, while the agent pauses when it needs a plan decision, creative clarification, review, or approval for credit-consuming work. Your MCP client may show the pause as a form or as a message in the conversation.

  • Examples: prompts for editing, generation, motion design, localization, and media enhancement.
  • Work with your files: local paths, chat attachments, public links, and project media.
  • Motion graphics: editable animation authored through your connected model.
  • Automation connector: the run lifecycle, approvals, continuation, and exports.