Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesShort answer: connect an MCP-compatible client to Lovable’s hosted endpoint, https://mcp.lovable.dev, authorize with OAuth, then describe the asset or app feature you want. The assistant can create or edit a Lovable project, attach source files, inspect the resulting code and files, and deploy the project. Lovable’s March 19, 2026 product update also documents generation of professional documents, PDFs, images and videos from the same conversation. Read-only inspection is free of workspace credits; agent actions that create or change things consume credits.
What Lovable MCP does
Model Context Protocol (MCP) is a standard way for an AI client to call tools in another service. Lovable hosts an MCP server at https://mcp.lovable.dev using Streamable HTTP and OAuth 2.1. After you authorize it, compatible clients such as Claude, ChatGPT, Cursor, VS Code, Codex and other MCP clients can ask Lovable to create, edit, inspect and deploy projects through natural-language requests.
The server exposes tools for workspaces and projects, agent messages, code and diffs, knowledge, databases, connectors, templates, analytics, file uploads and deployment. File-upload tools can create upload URLs, which lets you attach reference images or other source files to an agent conversation. Lovable’s March 19, 2026 Community Hub update says the product can analyze files and data, generate professional documents, and create images and videos. The formats named in that update include PowerPoint, Word, PDF, CSV, Excel, JSON, XML, images and video.
There is an important distinction: the MCP server is the control plane for Lovable projects. It is not documented as a standalone video-rendering, PDF-rendering or image-generation API with a fixed endpoint for each media type. You normally ask the Lovable agent to create the asset or to build an app feature that produces it, then inspect, download or deploy the result.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
Connect an MCP client to Lovable
- Choose a compatible client. Lovable’s tutorial names Claude, ChatGPT, Cursor, VS Code and Codex; Lovable is also available as an out-of-the-box MCP agent in Atlassian Rovo.
- Add the server endpoint. In the client’s MCP or connector settings, add
https://mcp.lovable.dev. The exact menu label varies by client; use its option for adding a remote Streamable HTTP server. - Complete OAuth. The browser sign-in grants Lovable access using your existing Lovable account and workspace permissions. The product page states that OAuth is the supported connection method and that there are no API keys to manage.
- Confirm the workspace. Ask the client to list your Lovable workspaces or projects. This is a read-only check and should not consume workspace credits.
- Start with a small request. Have the agent inspect a project or list files before asking it to create a project, generate an asset or deploy. You will see the proposed tool calls in the client and can approve them according to that client’s confirmation settings.
If the endpoint is not accepted, check that the client supports remote Streamable HTTP MCP servers and OAuth 2.1. A local-only MCP configuration or a client that expects a command to launch a local process cannot use this hosted endpoint as-is.
The basic Lovable build-and-deploy workflow
- List workspaces and projects. This establishes which account and workspace the client is using.
- Create or select a project. For a new project, state the product, target users, data sources and the media output you need.
- Send a precise build message. Include dimensions, page or document structure, branding, input files, output format and acceptance criteria. For example: “Create an invoice app that accepts a customer record and exports a two-page PDF with our logo, item table, tax and payment terms.”
- Inspect the response. Ask for the changed files, a diff, the data model and any connector configuration. Read-only code and analytics inspection does not use credits.
- Run a controlled revision. Request one change at a time, such as “add a landscape PDF option” or “replace the hero image with the attached file.” Each agent action that creates or changes content consumes workspace credits.
- Deploy when ready. Use the deployment tool to obtain a live URL, then test the generated document, image or video in the deployed experience.
Keep source material in the conversation or upload it through Lovable’s file tools. For images, the MCP server can generate an upload URL so the client can attach the file before the build message runs. For sensitive files, verify the workspace and connector permissions before uploading.
Generate a PDF or other document
Describe the document contract
Tell Lovable what enters the system, what the reader must see, and what file you expect. Include page size or orientation, margins, headers and footers, typography, tables, page-break rules, locale, currency and accessibility requirements. If the output is an invoice, specify tax treatment and rounding rather than leaving those decisions to the model.
Provide examples and data
Attach a representative document, spreadsheet or JSON sample. Ask the agent to identify fields and show a proposed schema before it rewrites the project. This reduces accidental changes when a source file contains merged cells, inconsistent dates or missing values.
Recommended Free Tools
Validate the generated file
Request a test record and inspect every page: clipping at the right edge, orphaned headings, unreadable tables, missing fonts, incorrect totals and links that do not open. If the PDF is produced by a deployed app, test the download in an incognito window as well as in the authenticated workspace.
Generate an image
Use an attached reference
Upload a logo, product photo or style reference through the file-upload tool’s generated URL, then state which parts may change and which must remain exact. For example, distinguish “use this logo without alteration” from “match this color palette.”
Specify production details
State the intended placement, aspect ratio, background, text-safe area, language and accessibility text. If the image is a UI asset, ask Lovable to place it in the project and expose an editable source or replacement path rather than embedding an opaque data URL.
Check rights and consistency
Review trademarks, faces, stock material and text for permission and accuracy. Generate variants only after the first composition meets the brief; otherwise each revision can consume additional credits without solving the underlying layout problem.
Generate a video or a video-producing feature
Choose the output you actually need
Ask for either a finished launch video or an app workflow that assembles clips, captions and branding. The March 19, 2026 update gives “create a launch video” as an example, but it does not publish a guaranteed duration, codec, rendering time or quality score. Treat those as requirements to test, not as service-level promises.
Supply a storyboard
Include the target duration, shot order, narration or on-screen text, aspect ratio, frame rate, music restrictions and the final delivery format. Attach source images, logos and copy. Ask the agent to produce a scene list first, then implement the generator or revise the asset.
Inspect before publishing
Check synchronization, spelling, safe areas for mobile crops, audio levels and whether the deployed app handles a failed or unusually large source file. If you need a repeatable pipeline, ask Lovable to expose the inputs and output location in the project rather than relying on a one-off chat result.
What costs credits, and which plans include MCP?
| Operation | Credit behavior | Practical implication |
|---|---|---|
| List projects, inspect files or diffs, check analytics | Read-only actions do not use workspace credits. | Use these checks to plan a change before spending credits. |
| Create a project, send an agent build message, edit code or data, deploy | Agent actions consume workspace credits. | Combine requirements into a clear request and make smaller revisions when possible. |
| Lovable MCP availability | Available on every plan, including Free. | No separate MCP subscription is required. |
| Enterprise workspaces | The product page says Enterprise customers should contact their account executive to enable MCP. | Activation may require an account-level step. |
Lovable’s published material does not state a universal credit price for a specific video, PDF or image operation. Credit consumption depends on the agent actions you request, so do not treat a media file as having a fixed published unit cost.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Lovable MCP versus publishing your app as an MCP server
| Question | Build with Lovable MCP | Publish a Lovable app as an MCP server |
|---|---|---|
| Who initiates the work? | A builder or internal team asks an AI client to drive Lovable. | Customers use your published app from ChatGPT, Claude or another assistant. |
| What is controlled? | Project creation, edits, files, data, connectors and deployment. | The tools derived from your app’s logic and the operations exposed to users. |
| Authentication | OAuth 2.1 with the user’s existing Lovable permissions. | OAuth is recommended; access can be limited to everyone, signed-in users or paying users. |
| Security controls | Workspace permissions govern what the connected client can see and change. | Choose the minimum tool scope and use read-only tools where possible; Lovable runs a security check when publishing. |
| Operations | You use Lovable’s hosted server while building. | Lovable hosts and updates the MCP server as your app evolves. |
Publishing is the route when the goal is customer-facing access to your app’s capabilities. Connecting to Lovable MCP is the route when the goal is to build, modify or deploy the app itself from an AI conversation.
Permissions, reliability and security practices
- Start with a read-only inspection and minimum necessary tool scope.
- Confirm the active workspace before uploading files or changing a database.
- Require review for deployment, connector changes and destructive edits.
- Use OAuth rather than attempting to invent or share an API key; the official product page describes OAuth as the supported method.
- Keep a copy of important source files and record the deployed URL after a successful release.
- Do not infer a quality, speed or success guarantee for generated media. The official sources reviewed do not publish controlled benchmarks, latency figures or success rates.
Common problems and fixes
The client cannot connect to the endpoint
Verify that you entered the exact endpoint, that the client supports remote Streamable HTTP, and that its OAuth browser flow completed. A client that only launches local MCP processes will need a compatible remote-server setting or a different client.
The agent can inspect but cannot build
Read-only tools may work while build actions are blocked by workspace permissions, a missing Enterprise enablement step or insufficient credits. Check the selected workspace and plan, then retry with an authorized account.
An upload is missing from the prompt
Generate a fresh upload URL through Lovable’s file-upload tool, attach the file in the same conversation and confirm that the agent can reference its filename before asking for the build.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →The output is not the requested format
Restate the required format and acceptance test in the build message, then inspect the changed files. For PDFs, specify page and typography rules; for video, specify container, duration and aspect ratio; for images, specify dimensions and background.
A deployment works for you but not for a customer
Check authentication and connector permissions in an incognito session. If the app is intended for external assistants, publish it as an MCP server and set the audience and tool scope explicitly instead of exposing internal workspace tools.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Or skip the browser setup
If you only need a clean image of a deployed Lovable app, ScreenshotNeo is a website screenshot API and MCP server. It accepts one request and returns PNG, JPEG, WebP or PDF. Cookie and consent banners, newsletter popups and chat widgets are removed before capture; bot checks, blank pages, timeouts, failed loads and cache hits are not billed, and response headers identify the page verdict and billing result. Its MCP tools—take_screenshot, get_page_info and capture_pdf—let Claude, Cursor or another MCP client capture pages without browser setup.
Use the API documentation at https://screenshotneo.com/docs/. A basic request is:
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://stripe.com -o shot.webpPython:
import requests
r = requests.get("https://api.screenshotneo.com/v1/shot", params={"access_key": "YOUR_API_KEY", "url": "https://stripe.com"}, timeout=90)
open("shot.webp", "wb").write(r.content)Node.js:
const q = new URLSearchParams({ access_key: 'YOUR_API_KEY', url: 'https://stripe.com' });
const res = await fetch(`https://api.screenshotneo.com/v1/shot?${q}`);The Free plan includes 1,000 screenshots per month with no card. Paid plans start at $5 for 3,000 shots, and every feature is included on every plan. Create a free ScreenshotNeo account.
FAQ
Can Lovable MCP be used from more than one AI client?
Yes. The same hosted endpoint is designed for MCP-compatible clients, including the clients named in Lovable’s setup material. Each client still requires its own connection and OAuth authorization flow.
Does publishing an app expose my whole Lovable workspace?
No. A published app exposes the tool scope derived from that app’s logic. Set the audience and minimum permissions deliberately, and prefer read-only operations when they are sufficient.
Where can I find a guaranteed rendering time or media quality score?
Lovable’s cited announcements do not publish controlled latency, quality or success-rate figures for video, PDF or image generation. Evaluate the result against your own files and acceptance criteria.
Bottom line
Use Lovable MCP when you want an AI client to build and operate a Lovable project, then ask that project to generate documents, images or video. Keep the two directions separate: connecting to Lovable builds the app, while publishing the app as an MCP server lets your customers use its selected capabilities. OAuth, least-privilege scopes, read-only inspection and explicit media specifications make the workflow safer and more predictable.
Frequently Asked Questions
Can Lovable MCP be used from more than one AI client?
Yes. The same hosted endpoint is designed for MCP-compatible clients, including the clients named in Lovable’s setup material. Each client still requires its own connection and OAuth authorization flow.
Does publishing an app expose my whole Lovable workspace?
No. A published app exposes the tool scope derived from that app’s logic. Set the audience and minimum permissions deliberately, and prefer read-only operations when they are sufficient.
Where can I find a guaranteed rendering time or media quality score?
Lovable’s cited announcements do not publish controlled latency, quality or success-rate figures for video, PDF or image generation. Evaluate the result against your own files and acceptance criteria.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
