Recommended Free Tools
Lovable’s official MCP server lets a compatible AI client work with your Lovable projects through natural-language requests. Lovable can also generate documents, images, and videos in the same conversation where you build a product. To use the two together, connect your AI client to Lovable, describe the asset or app feature you want, provide any relevant source files, then review the result and share or deploy it as needed. These are related capabilities, but they are not the same thing: MCP connects an AI client to Lovable, while Lovable’s asset-generation features produce the requested files.
What the Lovable MCP server does
The official Lovable MCP server is a hosted endpoint at https://mcp.lovable.dev. MCP, or Model Context Protocol, gives an AI client a standard way to ask another service to take actions. In this case, a connected client can work with Lovable workspaces and projects: for example, it can create or edit a project, inspect code changes, work with project data and integrations, and deploy an app.
The connection is useful when you want to describe a change in a conversation instead of navigating every project action yourself. The AI client sends a request through MCP; Lovable performs the corresponding action using your account’s access. Lovable’s MCP 101 tutorial, published August 18, 2026, describes MCP as a way for AI tools to interact with and take action in other tools. The official README says the hosted server uses Streamable HTTP and OAuth 2.1.
MCP does not mean that every kind of request is an MCP tool. It is the connection and action layer. Lovable’s Community Hub update of March 19, 2026 describes the separate ability to generate professional documents, images, and videos in the same conversation used to build a product.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →#1 Best Overall
What you can generate—and what the evidence establishes
Lovable’s March 19, 2026 Community Hub update lists PowerPoint, Word, PDF, CSV, Excel, JSON, and XML documents, as well as images and video. Its examples include creating a pitch deck or invoice and asking Lovable to make a launch video. You can request an asset directly, or ask for an app feature that creates or presents that kind of asset.
| Goal | What to ask for | What to review |
|---|---|---|
| Document or PDF | Describe the document, audience, content, and desired format; provide source data or files when relevant. | Check the figures, wording, page layout, and whether the result is the file you need or an app feature that produces it. |
| Image | Describe the subject, intended use, visual direction, and any required source material. | Inspect the image and confirm it is attached to the intended project or deliverable. |
| Video | Describe the purpose, audience, key message, and source material for the video. | Review the generated result before publishing or sharing it. |
| Feature in a Lovable app | Explain how a user should create, view, or download the asset within the app. | Inspect the project changes and test the relevant flow in the app before deploying. |
The Community Hub update is the direct source for the broader PDF, image, and video claims; the MCP README specifically documents upload URLs for attaching images. That upload capability is useful when the assistant needs image files as inputs, but it is not by itself a promise that every file format can be uploaded through the same tool.
Official sources reviewed for these capabilities do not publish a controlled benchmark, latency figure, quality score, or success rate for generated videos, PDFs, or images. The available information supports what formats and examples Lovable describes, not a quantified guarantee about output quality or completion time.
Connect an AI client and use Lovable for an asset
- Choose a compatible client. Lovable’s MCP 101 tutorial names Claude, ChatGPT, Cursor, VS Code, and Codex; the official README also gives setup examples for several MCP clients, including Claude Code, Claude Desktop, Cursor/Windsurf, VS Code, and Codex CLI. Atlassian Rovo is described as offering Lovable as an out-of-the-box MCP agent.
- Add the Lovable MCP endpoint. In the client’s MCP configuration or server setup, add
https://mcp.lovable.dev. Follow that client’s current setup instructions; configuration screens and file locations vary by client, so use its own current documentation rather than copying a configuration meant for another application. - Authenticate with Lovable. Complete the OAuth sign-in flow when prompted and authorize the account or workspace you intend to use. Lovable’s product page says the connection uses OAuth and does not require you to manage an API key. Access is governed by your existing Lovable permissions.
- Choose the task and supply context. Tell the assistant whether you want a standalone asset or an app feature. Give it the intended audience, purpose, content, preferred output format, and any constraints. Attach source material where appropriate; the MCP README documents tools that can generate upload URLs for images.
- Ask for the asset, then inspect it. For example: “Create a one-page PDF invoice from these details,” or “Create a launch video for this product using the attached images and this key message.” For an app feature, say how users should create or access the result. Review the returned file or the project changes rather than assuming the first draft is ready to publish.
- Deploy or share only when appropriate. If the task changed a Lovable app, use the assistant to inspect the diff and deploy when the project is ready. The practical MCP workflow in the README is to list workspaces, create a project, send a build message, inspect the diff, and deploy to obtain a live URL. A standalone generated file does not need an app deployment unless you want to share it through one.
Writing a brief that produces a useful first draft
Make the request specific enough to guide the result, but leave room for Lovable to build the format or flow. Include the source of truth for names, numbers, product details, and claims; identify who will use the result; specify the intended format; and call out any brand or accessibility requirements. For a video, add the main message and the role of supplied media. For a PDF, explain the document’s sections and the data that must appear. For an app feature, describe the user’s actions and the desired end state.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
When a source file matters, explicitly say whether the assistant should use it as factual input, visual reference, or both. That helps separate content instructions from design guidance. If a result has legal, financial, or other high-stakes content, verify it against the source material before using it.
Can Claude or ChatGPT build and deploy a Lovable app?
Yes, when the client supports MCP and is connected to Lovable with the required OAuth permissions. Lovable’s tutorial names Claude and ChatGPT among compatible clients, and the official README describes project creation, editing, diff inspection, and deployment workflows. The client can ask Lovable to make changes and, when appropriate, deploy the project to produce a live URL.
This is not the same as granting an AI client unrestricted access to every workspace. Lovable says the server uses the user’s existing permissions to limit workspace access. Review the OAuth authorization and confirm the selected account or workspace is the one you intend to use. You remain responsible for checking generated code, project changes, and the deployed result.
Lovable MCP credits, plans, and permissions
Lovable’s product page says MCP is available on every plan, including Free. Enterprise customers should contact their account executive to enable it for their workspace. The sources describe credit behavior by action rather than publishing a per-action price table:
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches- Read-only actions such as listing projects, inspecting files, and checking analytics do not use credits.
- Actions that ask the Lovable agent to create or change something consume workspace credits. Examples include creating a project or sending a chat message to the agent.
- The supplied official information does not state a universal number of credits for generating a particular PDF, image, or video. The cost depends on the workspace’s credit rules and the agent actions involved; check the workspace’s current plan and usage information before running substantial work.
In practice, separate inspection from creation when you can: use read-only requests to understand the project first, then give a focused change request. This makes it clearer which actions are likely to consume credits and avoids asking the agent to make unnecessary edits.
Lovable MCP versus publishing your app as an MCP server
These are two different directions of connection. With Lovable’s official MCP server, you—the builder—connect an AI client to Lovable so it can help create, inspect, change, or deploy a project. When you publish a Lovable app as an MCP server, users of that app can connect an AI assistant to the app’s capabilities.
| Question | Build with Lovable MCP | Publish an app as an MCP server |
|---|---|---|
| Who initiates actions? | Your AI client acts on your Lovable project. | An end user’s assistant calls tools exposed by your published app. |
| Typical audience | A builder or internal team working on a project. | Customers or other users who want to use app functionality from an assistant. |
| Access control | OAuth and your existing Lovable permissions govern project access. | The app owner chooses an audience such as everyone, signed-in users, or paying users; access defaults to OAuth-protected unless made public. |
| Who operates the MCP server? | Lovable hosts the official server that connects clients to Lovable. | Lovable hosts and updates the published app’s MCP server as the app evolves. |
| Typical result | Project changes, inspection, generated assets, or a deployed app. | App functions exposed as tools that an AI assistant can call. |
For a published app, Lovable can propose a tool scope based on the app’s logic. The owner can choose which tools to expose and who may use them. Lovable’s MCP 101 tutorial recommends starting with minimum access and using read-only tools where possible; it also says Lovable security-checks the server on publish. Treat tool scope as a product and security decision: expose only the actions users need, and be particularly cautious about tools that change data or trigger consequential actions.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Common problems and practical fixes
The client does not connect
Confirm the configured server address is exactly https://mcp.lovable.dev, that the client supports MCP, and that you completed the OAuth flow. If the client has separate setup paths for local and hosted servers, use the hosted-server path. The supplied Lovable information does not establish identical setup screens or configuration syntax across clients.
The assistant cannot see the intended workspace or project
Check which Lovable account completed OAuth and whether that account has access to the workspace. Existing Lovable permissions limit what the connection can do; reconnect with the intended account or ask a workspace administrator to grant the appropriate access.
A requested file or app change is missing
Clarify whether you asked for a standalone file or a feature in the project. Supply the source content or files the task depends on, state the format explicitly, and ask the assistant to inspect the project diff or identify the resulting file. An image upload URL documented by the README supports attaching images; it should not be treated as evidence that every asset type is handled through that upload mechanism.
You are unsure why a request used credits
Distinguish read-only inspection from an action that asks the Lovable agent to create or modify something. Listing or inspecting project information is described as credit-free, while agent actions such as creating a project or sending a build message consume workspace credits. Check current workspace usage for the details applicable to your plan.
The published app’s MCP tools expose too much
Revisit the proposed tool scope and access audience before publishing. Remove tools that are not required, prefer read-only actions where they meet the use case, and decide deliberately whether access should be public, limited to signed-in users, or limited to paying users. Lovable’s publish-time security check is useful, but it does not replace choosing least-privilege access for the app.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Or skip the browser setup
If your goal is to capture a screenshot or PDF snapshot of a Lovable page after it is deployed, ScreenshotNeo is an alternative to try first: it takes a website URL and returns a PNG, JPEG, WebP, or PDF. It is not a replacement for generating a video, image asset, invoice, or presentation. For a page capture, one GET request is enough:
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://stripe.com -o shot.webpSee the ScreenshotNeo API documentation for setup and options. Before capture, it accepts cookie or consent banners like a visitor and removes more than 60 known consent platforms, newsletter popups, and chat widgets; each step can be turned off. Bot checks, blank pages, timeouts, failed loads, and cache hits are not billed, and responses identify the page verdict and billing status in headers. Its MCP server provides take_screenshot, get_page_info, and capture_pdf tools for AI agents. The Free plan includes 1,000 shots a month without a card; paid plans start at $5 for 3,000 shots.
Sign up for ScreenshotNeo’s free plan: 1,000 screenshots a month, no card required.
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.
