Show HN: I am Building an Open-Source Confluence and Notion Alternative

If you are building an open-source Confluence and Notion alternative, one of the first hard problems is not the editor. It is moving knowledge around without destroying it.

I tested this recently while collecting notes from Confluence spaces, Notion pages, GitHub issues, docs sites, and a few private internal dashboards. The goal was simple: get the page into clean Markdown so I could paste it into ChatGPT, Claude, or Cursor without dragging along navigation chrome, broken tables, cookie banners, or a wall of useless text.

That sounds basic, but it matters. AI tools are much better when the input is structured. Headings, lists, links, tables, and code blocks give the model a map. A copied browser selection usually gives it mud.

This is where Web2MD fits.

Web2MD is a Chrome extension that converts the current web page into clean Markdown inside your browser. It is designed for people who want to use AI tools with real web content, including logged-in pages that server-side readers cannot reach. It has a free tier with 3 conversions per day, a built-in token counter, and one-click send-to-AI shortcuts for tools like ChatGPT, Claude, and Cursor. Pro is $9/mo.

It is not a full Confluence or Notion replacement. It does not host your wiki. It does not manage permissions. It does one smaller job: it turns the web page you are already looking at into Markdown that is usable by humans and AI systems.

For teams experimenting with open-source knowledge bases, that small job can unblock a lot.

Why Markdown still wins for AI workflows

Confluence and Notion are good at being apps. Markdown is good at being portable.

When I am evaluating a knowledge base, I usually care about a few practical questions:

• Can I export the page without losing structure?
• Can I diff it in Git?
• Can I paste it into an LLM without wasting half the context window?
• Can I keep code blocks, tables, links, and headings intact?
• Can I use it locally without sending private docs through another API?

Markdown is not perfect, but it answers those questions better than most rich-text formats.

That is why many open-source Confluence and Notion alternatives eventually orbit around Markdown, MDX, or a close cousin. The format is boring, inspectable, and easy to pipe into other tools.

The missing step is often extraction. You have a useful page in Confluence, Notion, Linear, GitHub, a private docs portal, or a customer dashboard. You want clean Markdown now, not after writing a scraper.

What I tested

I tested Web2MD on a mix of pages that are common in internal knowledge workflows:

• A public documentation page with nested headings and code blocks
• A logged-in Notion page
• A private Confluence page
• A GitHub issue thread
• A long marketing page with navigation, CTAs, and repeated footer links
• A dashboard page that a server-side fetcher could not access

The important result was not that every page became perfect. Some pages with heavy client-side rendering still needed a quick cleanup. But the useful pattern was consistent: because Web2MD runs in Chrome, it can read the page after I am logged in and after JavaScript has rendered the content.

That is the core difference from many server-side tools.

If a page requires a session cookie, SSO, a VPN, or a browser-only paywall flow, a remote reader often cannot see it. Your browser can. Web2MD works from that position.

Example: a messy knowledge-base page becomes usable Markdown

Here is a simplified version of the kind of output I want from an internal page:

# Product Analytics Migration Plan

## Goal

Move dashboard ownership from the legacy warehouse to the new event pipeline
without breaking weekly reporting.

## Systems involved

| System | Owner | Status |
| --- | --- | --- |
| Segment | Data Platform | Active |
| Snowflake | Analytics | Active |
| Looker | BI | Migration pending |

## Migration steps

1. Freeze new legacy dashboard creation.
2. Rebuild the top 12 weekly dashboards in Looker.
3. Compare 30 days of results between old and new sources.
4. Publish the new dashboard index.
5. Archive legacy dashboard links.

## Open questions

- Who owns backfill validation?
- Should archived dashboards redirect to the new index?
- Do customer-facing reports need a separate QA pass?

That is the kind of Markdown I can paste into Claude and ask for a risk review. It is also the kind of Markdown I can commit into an open-source docs repo, split into smaller files, or use as seed content for a Confluence or Notion alternative.

The token counter matters here. Long internal pages can look harmless until you paste them into an AI tool and burn most of the context window. Web2MD shows an estimated token count before sending the content, which helps me decide whether to paste the whole page or summarize sections first.

How Web2MD compares to Jina Reader, Firecrawl, and MarkDownload

There are good tools in this space, and I do not think one tool wins every case.

Jina Reader is excellent when you want a fast server-side reader URL for public pages. It is simple, scriptable, and useful for public web research. If the page is reachable from the public internet, Jina Reader is often the quickest path.

Firecrawl is strong for crawling and extraction at scale. If you need to crawl a whole site, build a dataset, or feed a pipeline, Firecrawl is closer to infrastructure than a clipboard tool.

MarkDownload is a useful browser extension for saving pages as Markdown. It has been around for a while and is a good option for people who want local Markdown capture from the browser.

Web2MD overlaps with these tools, but the edge is specific:

• It runs browser-side, so it works on authenticated pages your browser can already see.
• It keeps the conversion local and private.
• It includes a token counter for AI context planning.
• It has one-click send-to-AI flows.
• The free tier does not require an API key and gives 3 conversions per day.

That makes it useful for knowledge workers, founders, developers, and researchers who are not trying to crawl the web. They are trying to convert the one page in front of them into something useful.

Why this matters for open-source Confluence and Notion alternatives

If you are building or evaluating an open-source alternative to Confluence or Notion, you probably have a migration problem.

Your team already has scattered knowledge:

• Some pages in Notion
• Some spaces in Confluence
• Some decisions in GitHub issues
• Some specs in Google Docs
• Some runbooks in private portals
• Some notes buried in product pages and dashboards

A perfect migration tool rarely exists because each system has its own structure and permissions. But a browser-side Markdown converter gives you a practical bridge.

You can open a page, convert it, inspect the Markdown, trim what you do not need, and move it into your new system.

Here is a second example, this time from a product requirements page:

# Requirements: Customer Export API

## Summary

Customers need a self-serve way to export account-level activity data for
internal reporting and compliance reviews.

## Non-goals

- Replacing the existing admin dashboard
- Supporting real-time streaming exports
- Building custom reports for each customer

## API shape

POST /exports

Request body:

```json
{
  "account_id": "acct_123",
  "start_date": "2026-07-01",
  "end_date": "2026-07-20",
  "format": "csv"
}

Acceptance criteria

• Export jobs are visible in the admin dashboard.
• Completed exports expire after 7 days.
• Failed jobs include a human-readable error message.

That output is not just easier to read. It is easier to review, version, summarize, and transform.

For example, you can paste it into Cursor and ask for implementation tasks. You can ask ChatGPT to find missing edge cases. You can ask Claude to turn it into a customer-facing changelog. Or you can commit it directly into a docs repository.

## Limits I noticed

Web2MD is not magic, and it is better to be clear about the boundaries.

First, it is Chrome-only right now. If your workflow is Firefox or Safari, that is a real limitation.

Second, the free tier is capped at 3 conversions per day. That is enough for occasional use, testing, and lightweight workflows. If you are converting a lot of pages, Pro is $9/mo.

Third, the quality of output still depends on the source page. Clean semantic HTML converts better than heavily nested app layouts. Some pages still need manual cleanup, especially if the original page uses complex widgets, hidden tabs, or unusual table rendering.

Fourth, it is not a crawler. If you need to archive an entire public website, use a crawling tool. Web2MD is best when you are working page by page from the browser.

Those limits are acceptable for my use case because the hard pages were not public websites. They were authenticated pages that normal server-side tools could not reach.

## A practical workflow

The workflow I liked most was simple:

1. Open the Confluence, Notion, GitHub, or docs page in Chrome.
2. Use Web2MD to convert the page.
3. Check the token count.
4. Copy the Markdown or send it to an AI tool.
5. Ask the AI tool to summarize, clean, split, or turn it into migration-ready docs.
6. Save the result in your new Markdown-based knowledge base.

If you are building a docs system, you can also pair this with your own import pipeline. Web2MD gives you a clean starting point. Your system can handle frontmatter, slugs, ownership metadata, review status, or whatever your knowledge base requires.

For more details on the extension itself, see the Web2MD homepage at [web2md.org](https://web2md.org). If you are comparing page-to-Markdown workflows, the Web2MD blog also has related notes on using Markdown with AI tools.

## Bottom line

The phrase "open-source Confluence and Notion alternative" can mean many things: a wiki, a docs repo, a collaborative editor, a knowledge graph, or a company memory system.

But almost all of those projects need better input. They need a way to turn existing pages into structured, portable text.

Web2MD is useful because it meets knowledge where it already lives: in your browser, behind your login, rendered the way you see it. It converts that page into Markdown, shows the token cost, and makes it easy to move the content into an AI tool or a Markdown-based system.

It will not replace your wiki. It will help you feed it.

If you are testing a Confluence or Notion alternative, try converting a few real internal pages with Web2MD and see how much cleanup you still need. The free tier gives you 3 conversions per day, no API key required.