Public keys will not work. Keys starting with
pk_ are meant for browser widgets and are rejected by every component in this bundle. Read more in
Private Keys.
---
## Perceive a page
**Perceive URL** takes the URL from the flow and two settings from the form:
| Parameter | Set by | Default | Purpose |
|-----------|--------|---------|---------|
| URL | The flow | -- | The absolute URL to render |
| Outputs | The form | `markdown` | Any of `markdown`, `html_cleaned`, `html_raw`, `screenshot`, `screenshot_full_page`, `pdf`, `links`, `images`, `structured`, comma-separated |
| Only Main Content | The form | Off | Strip navigation, headers and footers, keeping the article |
The component returns each artifact as a signed URL that expires in 15 minutes, plus the text of markdown, cleaned HTML and raw HTML inlined up to 256 KB each. Anything larger, and anything binary, stays a link. The `structured` output is the exception, arriving inline.
---
## The render_quality score
Every Perceive result carries `render_quality`, a number from 0.0 to 1.0. Zero means the page was blocked or empty; one means a clean render. The content still comes back either way, flagged, so a bad read never quietly enters a summary.
Use an **If/Else** component after Perceive to branch on the score before passing the markdown to an LLM component. A low score is a signal that the page is not what the agent expected.
---
## Searching and discovering
**Web Search** takes a query and optional filters and returns an array of results, each with title, URL, snippet and position.
**Discover URLs** takes a site URL and a mode (sitemap, crawl, or hybrid) and returns a flat list of URLs on that site without rendering each one. Use this before Perceive to map a site when you want to know what's there without the rendering cost.
---
## Extracting fields
**Extract Structured** needs a schema and exactly one source. The schema is a JSON object in one of two shapes. A flat map of field name to plain-language description is the quick one:
```json
{"title": "the page title", "price": "the monthly price in USD"}
```
A full JSON schema, `{"type": "object", "properties": {...}}`, is the other, worth the typing when you need types or nested objects enforced.
For the source, pass **URLs**, a comma-separated list capped at 50, or **Discover From URL**, a single site the API crawls for you, shaped by **Discover Mode** and **Discover Max Pages**. Passing both, or neither, returns an error.
Results arrive as an array, one entry per URL, each with its own data and its own `extraction_tier`, which says whether a deterministic pass or a model reading the page produced the answer.
---
## Converting files
**Convert File to Markdown** and **Convert File to PDF** take a Langflow **file**, not a URL. Wire in a file from an **Input** component's file field, or a file variable from an earlier component. The component uploads the bytes and returns a signed download URL for the result, valid for about 15 minutes.
The uploaded filename's extension decides the input format. Langflow carries the original name through, so this is normally invisible, but a file arriving without a usable extension falls back to its own bytes: a PDF, an image or a DOCX is recovered from its signature, while a text format such as CSV or HTML has none and is rejected with a 400.
---
## Building with it
### An agent that researches
Give an agent **Web Search**, **Perceive URL** and **Discover URLs**. Search finds candidate pages, Perceive reads the promising ones as markdown, and Discover maps a site when you need to know what else is on it without rendering anything.
### A workflow that extracts structured data
Chain **Discover URLs** into **Extract Structured** to pull typed fields off every page on a site at once, crawled and extracted server-side, delivered as one results array.
### A workflow that turns an upload into a summary
```text
Input (file) → Convert File to Markdown → Text Splitter → LLM → Answer
```
Convert the file to markdown, split it into chunks, pass each chunk to an LLM for summarization, and combine the results.
---
## Troubleshooting
**`The component fails with a 401 or 403 error`.**
The API key is wrong, missing, or it is a public `pk_` key. Verify you stored a **private** key (`sk_` prefix) in the global variable and selected it in the component's **EnConvert API Key** field. Test the key in the [dashboard](/dashboard/api-keys) by clicking its row.
**`Perceive returns a link where markdown was expected`.**
The markdown exceeded the 256 KB inline limit, so the component handed over the signed URL instead. Fetch it with a plain GET and no API key. **Only Main Content** usually brings a long page back under the limit.
**`A component reports the operation is not in my plan`.**
Perceive, Web Search, Discover and Extract Structured draw from your plan's monthly ops allowance. Check your usage on the [pricing page](/pricing.md).
**`Convert File to Markdown returns a 400 with "unsupported format"`.**
The file has no recognizable extension and its bytes don't match a known file signature. Add or fix the filename's extension (e.g. `.pdf`, `.docx`, `.csv`) and try again.
---
## Source and links
- **Source**: [enconvert/langflow-enconvert](https://github.com/enconvert/langflow-enconvert)
- **Package**: [PyPI](https://pypi.org/project/lfx-enconvert)
- **Licence**: MIT
- **Langflow documentation**: [docs.langflow.org](https://docs.langflow.org)
- **Underlying API**: [Introduction](/docs/introduction.md)
---
## Frequently asked questions
### How do I install the EnConvert bundle in Langflow?
Install the PyPI package (`pip install lfx-enconvert`) and restart Langflow. The six components appear under the **EnConvert** bundle in the sidebar. No config needed.
### Can I use a public API key?
No. Public `pk_` keys are for browser widgets only and are rejected by every component. Generate a private key in the [dashboard](/dashboard/api-keys) and use that.
### How do I store the API key so it doesn't appear in the flow JSON?
Create a **global variable** in Langflow (**Settings > Global Variables**), give it your private key, and select that variable in each component's **EnConvert API Key** field. Global variables are stored securely and never exported in the flow.
### How do I perceive a web page in Langflow?
Add the **Perceive URL** component, pass it a URL, and leave **Outputs** at its default. The component renders the page and returns the markdown inline in the component's output. Turn on **Only Main Content** to drop navigation and footers.
### Can I extract structured data from multiple pages at once?
Yes. **Discover URLs** lists all URLs on a site, and **Extract Structured** accepts a comma-separated list of URLs (up to 50). Pass the discovered URLs and the schema into Extract Structured to pull the same fields off every page.
### What does render_quality mean?
Every page Perceive renders gets a `render_quality` score from 0.0 (blocked or empty) to 1.0 (clean render). Branch on this score with an **If/Else** component before passing the content to an LLM, so your agent distrusts a low-quality read.