The file backend behind CourseLit and SendLit
Three ways in, one set of files
An agent saves a report. You check it in the dashboard. Your app serves it from the same URL. Nobody copies files between tools.
Adding uploads to your app? Skip the plumbing
Storing user files in S3 yourself means presigned URLs, CORS rules, thumbnails, a database of what you stored and a job to clean up what nobody uses. MediaLit already does all of that, so you write one API call instead.
Direct browser uploads
Your server hands out a short-lived signature and the browser uploads directly. Your API key never leaves your server.
No orphaned files
Uploads stay temporary until your app decides to keep them. Anything nobody keeps is deleted after 24 hours, so abandoned uploads never pile up.
Public and private files
Public files get a permanent URL. Private files are served through signed URLs that expire.
Resumable uploads
Large files upload in chunks and pick up where they left off after a dropped connection.
Thumbnails
Thumbnails are generated for images and videos on upload, with optional WebP conversion.
Every file, listed
List your files with their details, paginated and filtered by group or access, without keeping your own database.
Up and running in a minute
Connect an agent, upload from your terminal or call it from your code. Same account, same files.
1. Connect an agent
Add the MCP server in Claude, ChatGPT, Cursor or any MCP client, then sign in. In Claude Code:
claude mcp add --transport http medialit https://api.medialit.cloud/mcp2. Upload from your terminal
Install the CLI, sign in once, and upload. In a CI job, set MEDIALIT_API_KEY instead of logging in.
npm install -g @medialit/cli
medialit login
medialit upload photo.jpg --public3. Call it from your code
Use the Node.js SDK on your server, or the REST API from any language.
const medialit = new MediaLit({ apiKey: process.env.MEDIALIT_API_KEY });
const media = await medialit.upload("./photo.jpg", { access: "public" });Simple pricing
Start free. Upgrade when you need more room. Every plan gets the API, the MCP server, the CLI and the dashboard.
Questions, answered
Is MediaLit just a wrapper around object storage?
No. Storing bytes is the easy part. MediaLit adds what you would otherwise build around it: browser uploads without exposing keys, resumable uploads, automatic cleanup of unused uploads, public and private access, a list of your files with their details, thumbnails, quotas, an MCP server for agents, a dashboard and a CLI.
Can I upload from the browser without exposing my API key?
Yes. Your server creates a short-lived upload signature and the browser uploads with it. The React uploader handles this for you.
Which AI assistants work with MediaLit?
Any client that supports remote MCP servers, including Claude, ChatGPT, Cursor and VS Code. You sign in with OAuth, so there is no API key to paste.
Can I upload from the terminal, a CI job or a coding agent?
Yes, with the CLI. Run medialit login on your machine, and you and any coding agent running there can use it. In CI, set MEDIALIT_API_KEY instead of logging in.
Can I use my own S3 or R2 bucket?
Yes, by self-hosting MediaLit on AWS S3, Cloudflare R2, MinIO or any S3-compatible storage. The hosted service at medialit.cloud stores files in its own storage.
Is the open-source version limited?
No. Everything that runs medialit.cloud is in the GitHub repository, licensed under AGPL-3.0.
What happens when I run out of storage?
New uploads are refused until you free up space or upgrade to Pro. Files you already have stay online and keep their URLs.
Who can see my files?
Only you, unless you make a file public. Public files get a permanent URL that anyone with the link can open. Private files can only be opened through signed URLs that expire. An agent gets access only after you sign in and connect it.