Onlora
Connect an agent

The creative layer for AI agents.

Characters. Images. Voices. A studio for your agent.

Your agent makes, animates and finds them. You watch it all on one canvas.

Connect an agentYour projects
A project canvas: a library photo of a lighthouse, a night version made from it, a cat character sheet, and two comic panels made from both, joined by lines.

One image. A character that moves.

Your agent cuts a drawing into layers. It blinks, reacts and talks, in a voice made from a description.

Motion
Calm, Lively or Shy.
Expression and reaction
Tap one. The face changes, or the whole character reacts, with a mark.
Voice
Describe a voice, or pick one and hear it in Japanese, Chinese and English first.
Clip
Record 3, 5 or 10 seconds on a clear, white, dark or green background.

The tools

Images use quota; speech, character clips and search are free.

Generate

  • generate_image
  • edit_image
  • generate_voice
  • design_voice

Characters

  • make_character
  • split_layers
  • speak_character
  • record_character

Search

  • search_library
  • search_comics
  • add_asset
  • crop_asset

Projects

  • create_project
  • get_project
  • write_file
  • create_share

Listen and check

  • analyze_audio
  • get_job
  • list_models
  • get_quota

The canvas

One generated panel opened: the image, the two references it used, its prompt and model.
The same project on a phone, as a list.
A comic open in the reader on a computer: page 2 of 14, with arrows on both sides.
  1. Provenance chain
  2. Prompt, model, license
  3. Replay
  4. Phone list
  5. Reading mode

Share

Send a link to a canvas, a character, a file, a book or a web page. Anyone opens it on any device, with no account.

A shared canvas opened in a browser without signing in: the cards, a Read button for a book, and a Made with Onlora link.
The same link on a phone, as a list.

The library

One search across images, sounds, video, fonts, comics, poses, costumes and voices. Every file keeps its license and credit, even after edits.

The library page: one search for pagoda with images, sounds and video in rows, each result with its license, source page and an Add to project button.
Open the library

Connect

Add this address to any MCP client that supports OAuth.

https://api.onlora.ai/mcp

Steps for each client·Onlora is invite-only for now.