Supertrove Public

SuperTrove Features & Use Cases

Everything SuperTrove can do — features, capabilities, and real-world use cases for AI agents and teams.

6 items

Semantic Search & Vectorize

featuresearchaivectorize

Summary

Every item is embedded automatically. search_list finds semantically similar items even when exact keywords don't match.

Content

## What it does

When an item is added or updated, SuperTrove generates a text embedding via Workers AI and stores it in Cloudflare Vectorize. `search_list` performs approximate nearest-neighbour search over those embeddings.

```mermaid
flowchart LR
    A[New Item] --> B[Workers AI\nEmbedding]
    B --> C[Vectorize Index]
    D([search_list query]) --> E[Embed query]
    E --> F[ANN Search]
    F --> G[Top-K Items]

When to use it

query_list — exact field matching, fast, no AI
search_list — natural language, finds related items by meaning

Example

search_list({ list_id: "...", query: "products for sensitive skin" })
→ returns Hydrating Serum, Gentle Cleanser, Barrier Cream
  even if none contain the phrase "sensitive skin" verbatim

MCP Protocol — Native Agent Integration

featuremcpagentsintegration

Summary

SuperTrove speaks MCP natively. Claude, Cursor, and any MCP-compatible agent can read and write lists without custom code.

Content

## What it does

SuperTrove exposes a full MCP endpoint at `https://api.supertrove.ai/`. Any agent that supports MCP can connect with a Bearer token and immediately use all tools.

```mermaid
sequenceDiagram
    participant C as Claude / Cursor
    participant M as SuperTrove MCP
    participant D as Durable Object
    C->>M: tools/call add_item
    M->>D: write JSON blob
    D-->>M: item_id
    M-->>C: { item_id }

Available MCP Tools

create_workspace / list_workspaces
create_list / list_lists
add_item / bulk_add_items
query_list / search_list / get_list
update_item / delete_item
get_public_list — read public lists without auth
share_workspace — invite by email

Connect in 30 seconds

Go to supertrove.ai/settings → generate an API key
Add to your MCP client: https://api.supertrove.ai/ with Bearer token
Done — your agent can now store and retrieve structured data

Schemaless JSON Storage

featurestoragecoreagents

Summary

Store any JSON shape in a named list — no schema definition, no migrations, no friction.

Content

## What it does

Every item in SuperTrove is a free-form JSON object. There's no schema to define upfront — you just push data and SuperTrove stores it.

```mermaid
flowchart LR
    A([Agent]) -->|add_item| B[SuperTrove List]
    B --> C[(Durable Object\nSQLite)]
    C --> D[D1 Index]

Why it matters

AI agents rarely know the shape of their output in advance. SuperTrove lets agents store whatever they produce — product enrichments, research notes, task queues, extracted entities — without a human configuring a database first.

Example

{
  "product": "Hydrating Serum",
  "sku": "HFS-50ML",
  "tags": ["skincare", "bestseller"],
  "seo_title": "Hydrating Face Serum – 50ml"
}

No migration needed. Add a new field tomorrow and it just works.

Public Lists — Share Without an Account

featuresharingpublicdiscovery

Summary

Set any list to Linked or Indexed. Anyone with the URL (or the slug) can read it — no login required.

Content

## What it does

Lists have three visibility levels:

| Visibility | Who can read | Discoverable |
|---|---|---|
| Private | Workspace members only | No |
| Linked | Anyone with the URL | No |
| Indexed | Anyone | Yes — appears in /explore |

## For agents

Public lists are accessible via the `get_public_list` MCP tool using only the list's slug — no API key needed by default, but the calling agent still uses its own token.

```mermaid
flowchart LR
    A([Any Agent]) -->|get_public_list slug| B[SuperTrove]
    B --> C{visibility?}
    C -- linked/indexed --> D[Return items]
    C -- private --> E[404]

Use case

Publish a shared knowledge base, product catalogue, or feature list that any team member's agent can pull on demand — without adding them to your workspace.

Collections & Knowledge Graph

featurecollectionsgraphorganisation

Summary

Group lists into collections. Link related lists together into a graph for structured knowledge navigation.

Content

## What it does

Collections let you group lists thematically. The knowledge graph lets you create explicit relationships between lists — so agents can traverse related content.

```mermaid
flowchart TD
    COL[Collection: Skincare Launch]
    COL --> L1[List: Product Data]
    COL --> L2[List: Content Briefs]
    COL --> L3[List: Campaign Assets]
    L1 -- related_to --> L2
    L2 -- related_to --> L3

Use case

An orchestrator agent queries the graph to discover all lists related to a campaign, then pulls content from each — without being hardcoded with list IDs.

Multi-Tenant Workspaces

featurecollaborationteamsaccess-control

Summary

Organise lists into workspaces. Invite team members with owner, member, or guest roles.

Content

## What it does

Workspaces are isolated containers. Each workspace has its own lists, members, and API keys. Guest access can be scoped to individual lists.

```mermaid
flowchart TD
    W[Workspace] --> L1[List: Products]
    W --> L2[List: Blog Ideas]
    W --> L3[List: Agent Logs]
    W --> M1[Owner]
    W --> M2[Member]
    W --> M3[Guest — List A only]

Roles

Role	Can create lists	Can invite	Can view all lists
Owner	✅	✅	✅
Member	✅	✅	✅
Guest	❌	❌	Assigned lists only

Use case

Invite a client as a Guest to only the list you want them to see. They can read items without access to the rest of your workspace.

Shared via SuperTrove