From members-only to public knowledge — why Bindly opened up
The product decision behind Bindly v1's public spaces: why the original members-only model was the wrong starting point, what changes when knowledge can be discovered and indexed, and what this means for how LLMs find and use Bindly content.
- v0 was private by default — all spaces were hidden, no public content, no SEO
# From members-only to public knowledge
Bindly v0 was invisible by default. Every Space was private. Every Binding was hidden behind authentication. If you hadn't installed the MCP extension, generated a user key, and loaded your personal space with content — Bindly didn't exist for you.
That was the right choice for a private note-taking tool. It was the wrong choice for a knowledge platform.
## The problem with invisible knowledge
The private-by-default model had a fundamental flaw: it required users to already know what they were looking for.
To use Bindly's MCP tools, you had to:
1. Install the MCP server in your LLM client
2. Authenticate with your credentials
3. Search for content *you had already created*
The MCP interface was a retrieval layer for your own notes. Useful, but limited. You couldn't point someone to a Bindly document. You couldn't share a link that just worked. Google couldn't index anything. Claude (without MCP) couldn't find anything. The knowledge existed only for the people who had already put it there.
The use case that kept coming up in practice: "I want to write something down and have it accessible later — from any LLM, without setup, if I want it to be."
That use case doesn't require private-by-default. It requires *choice*.
## What changed in v1
v1 introduces three Space types with different default visibility:
- **Personal**: always private. Your personal notes, your private knowledge. Never indexed, never public.
- **Team**: private by default, but toggleable. Flip it to public and the whole Space becomes accessible at `bind.ly/@yourspace`. Flip it back, private again. This is the GitHub repository model.
- **Official**: always public. Created and managed by Bindly. `@bindly`, `@bindly-feedback`. Permanently public, verified, prioritized in llms.txt and sitemaps.
Public Team and Official spaces get permanent URLs for every Binding:
```text
bind.ly/@bindly/cloudflare-native-stack
bind.ly/@bindly/mcp-tools-tier-system
```
These URLs are indexed by Google. They're in the sitemap. They have OG tags and JSON-LD structured data. They work for anyone — logged in or not, MCP-connected or not.
## The Markdown URL as a universal access layer
The most significant change for LLM access isn't the OAuth flow or the MCP tools — it's the `?format=md` convention:
```text
bind.ly/@bindly/cloudflare-native-stack?format=md
```
That URL returns clean Markdown. No HTML to parse. No authentication required for public content. No MCP setup.
Claude with web access can read any public Bindly document directly. A Python script can fetch it with `requests.get()`. A GitHub Action can pull it as part of a workflow. The MCP tools are the *convenient* interface — `?format=md` is the *universal* one.
The `textUrl` field in every MCP response surfaces this automatically:
```json
{
"textUrl": "https://bind.ly/@bindly/cloudflare-native-stack?format=md"
}
```
When an LLM creates or retrieves a Binding via MCP, it always gets the public URL alongside the content. It can include that URL in a response, and anyone can access it later without any MCP setup.
## What this does to the knowledge model
Private-by-default Bindly was a personal knowledge tool. Public-capable Bindly is closer to a knowledge publication platform — one that happens to have excellent LLM access.
**Private workflow**: Claude reads your meeting notes, extracts decisions, creates a Binding. You retrieve it later via MCP when you need context for the next meeting. Nobody else sees any of this.
**Public workflow**: You write a technical note about a decision your team made. You publish it to your public Team Space. Your teammate reads it at `bind.ly/@yourteam/that-decision`. Claude (for anyone on the team) can search Bindly for it. Google indexes it. It becomes a document that exists in the world.
**Official workflow**: Bindly publishes documentation and guides to `@bindly`. LLM crawlers index it. Users searching for "how does Bindly MCP work" find the answer. The knowledge becomes discoverable without anyone needing to know to look in Bindly specifically.
## The auto-accept member model
Public Spaces use what we think of as the "public café" model: joining is automatic. When a Team Space is public, any authenticated user can become a member by clicking "Join." No approval required. This is the same pattern as GitHub's public repositories — you can open a PR without being a collaborator first.
Private Spaces work differently: joining requires either an invitation from an admin or an explicit approval process. The distinction maps naturally to what users expect from private vs public context.
The permission model within a Space doesn't change based on public/private status. Members create and edit their own content. Admins can edit or remove anyone's content (important for public Spaces where spam is a risk). Owners have full control.
Rate limits apply at the API level — 10 bindings per minute per user, 5 public Team Spaces per account — to prevent abuse of the auto-accept model.
## What didn't change
The personal Space is still completely private. There's no way to make a personal Space public — it's built into the type. If you want to publish something, you create a Team Space and toggle it public.
Individual Binding visibility doesn't exist. The Space's visibility setting controls everything. If the Space is public, all its Bindings are public. This is simpler to reason about and avoids the common mistake of forgetting to make a specific document public when you intended to.
Private content you want to share temporarily still works via share links (`bind.ly/b/{shareId}`) — unlisted, not indexed, with optional expiry. A middle ground between fully public and fully private.
## The bet
The bet in v1 is that Bindly's value isn't just in private retrieval — it's in *knowledge that persists and is findable*. Public spaces mean content can be discovered by people who didn't know to look for it. Markdown URLs mean LLMs can access content without any integration work. Indexing means the knowledge compounds over time.
Private-by-default felt safe. Public-capable is more honest about what a knowledge platform should do.