Lorediv is a public index of expertise: readable, structured data, so agents can position you. It is an open knowledge tree where topics, branches and sources carry the author's lore rank.

How does Lorediv rank expertise?

Rank comes from community validation, not self-declaration. It grows by creating topics and by contributing to the community's trees. Every source arrives with the track record behind it.

Can AI agents read Lorediv?

Yes. Every page is published as HTML for people and as Markdown and JSON for agents, at the same URL with a .md or .json suffix. There is no login wall and no scraping required.

Is Lorediv open source?

The formulations that decide rank, energy, votes, reputation, trending, community and tags are published under AGPL-3.0. The interface and this landing page are not part of the public repository.

How do I join the beta?

The beta opens soon. Access is requested through the Lorediv Discord at https://discord.gg/GFBsQQ7pCC.

Lorediv — The Open Expertise Index

Gather · Structure · Connect

Scattered to
connected expertise

Your work, everywhere

Your valuable content is scattered across the web, different platforms, different handles, with no clear structure. AI agents can’t identify your expertise or cite you as a trusted voice.

Build your Lorediv site, compound your content, and let your public lore speak for you.

Wiki/lorediv_wiki Core idea/core_idea

x.com/@priya_nair

Notes on open models

POST

github.com/priya-nair

Inference tools & code

CODE

priya-nair.substack.com

Research & essays

ESSAY

arxiv.org/a/nair_p_1

Quantization research

PAPER

youtube.com/@priyanair

Talks on local inference

TALK

kaggle.com/priya_nair

Open model benchmarks

DATA

Lore

6Topics

14Branches

57Received

9Active days

JanFebMarAprMayJunJulAugSepOctNovDec

Specialization DivsTop 5

Structured for discovery. Built for an open web.

Grok

Anthropic

GPT

Perplexity

Gemini

Tell the agentic web who you are

Make your content
position you

Lorediv makes your expertise findable: structured, public, connected to keywords, and growing as context. It gives your work a permanent address, a place in an open knowledge tree, and a clear signal for agents like Grok, Claude, ChatGPT, Perplexity and Gemini, so next time they can cite you in your field.

TopicsUsersLore

Run local models on an M4 MacBook Prolorediv.com/@priya_nair

SCHDM+5 placed +42 branches

GOAT 8.4/localai/nvidia/models

Quantized models, testedlorediv.com/@wei_zhang

CHDMA+8 placed +58 branches

Wizard 6.1/llms/localai/nvidia

Top 5 local LLM tools comparedlorediv.com/@noor_ali

HDMAK+11 placed +19 branches

Geek 3.5/rag/llms/pytorch

Setup notes for local LLMslorediv.com/@mei_lin

DMAKL+14 placed +27 branches

Wizard 6.3/agents/grok/claude

Top 5 of 1,204 users placedIllustrative data

Be the source

Search agents don’t browse. They pick sources, and they pick before they read. Lorediv shows them who really knows a topic, so your expertise is what they cite.

Document twins/agents/document-twins Search agents/rank#search-agents What is Lorediv/getting-started/what-is-lorediv

agent session · one question

One tree, every format

Knowledge doesn’t arrive finished. It emerges along the way. A question, a claim, a proposal, a prediction? Pick the shape that fits, and watch the tree take form.

Topic types/topic-types Post/topic-types/post Ask/topic-types/ask

/post

Share an idea. Responses nest as letters.

/ask

Ask a question and rank the answers.

/proposal

Put motions up. One +1 each, and it can move.

/poll

A short vote that closes in 7, 13 or 30 days.

/prediction

State a claim with a close date. Counters compete.

/context

Wrap outside work, keep it attached to its author.

/argument

Write Argdown and get the map drawn.

/graph

Draw a flow with nodes and edges.

index · /sources

/sources
├─ /post Field notes on public rank
├─ /context Thread by @ada
│   ├─ embed x.com/…/status
│   └─ A · Open the three sources
├─ /context Paper on benchmarks
│   └─ embed arxiv.org/abs/…
├─ /context Eval harness
│   └─ embed github.com/…
├─ /proposal Cite closed topics
├─ /poll Which source first?
├─ /prediction MCP by 2027
├─ /argument Where the benchmark leaks
└─ /graph Agent read path

JSON

{
  // Topic type (REQUIRED)
  "topic_type": "proposal",

  // Title: the debate, not a clone of branch A
  "title": "How should a founder run the first enterprise motion?",

  // Framing line (REQUIRED, unique)
  "subtitle": "Jen Abel's 15-step cycle vs the five CRM stages founders still copy into HubSpot.",

  "tags": ["gtm", "enterprise", "sales"],

  "branches": [
    { "body": "Pincer the exec and the n-1; withhold the demo until the champion co-authors it" },
    { "body": "Keep five CRM stages and demo on the first qualified call" },
    { "body": "Time-box a 2–3 day pilot with 3–4 power users before any price talk" }
  ],

  "expires_at": "90d"
}

From readable, to answerable, to reachable

Humans keep talking to humans in every phase. What changes is how agents meet the tree, and how they reach you.

Findable

Phase 1

Every public page is a document. No session, no key, no scraping: the tree ships as HTML, Markdown and JSON at the same address.

# no session, no key
GET /topics/post/local-models.json

{
  "schema_version": 1,
  "author": {
    "handle": "@priya_nair",
    "rank":   "Wizard 6.1",
    "score":  6.1,
    "divs":   ["localai"]
  },
  "relevance": 0.91,
  "votes":     42
}

Queryable

Phase 2

REST and MCP read the same index. An agent asks a question and gets ranked, attributed sources it can cite, not a page it has to parse.

query({
  "tag":      "quantization",
  "min_rank": "wizard",
  "cite":     true
})

→ [{
  "author":  "@wei_zhang",
  "rank":    "GOAT 8.7",
  "relevance": 0.94,
  "cited_by": 37,
  "url":  "/post/4-bit-at-8k"
}]

Reachable

Phase 3

You never know when an agent will need a human. Phase 3 opens a mailbox: inquiries reach you there, and their answers stay yours.

POST /agents/message
{
  "to":      "@mei_lin",
  "from":    "diligence-agent",
  "on":      "/post/enterprise-motion",
  "purpose": "expert_review",
  "body":    "Does this pilot"
             "hold for"
             "regulated buyers?"
}

→ 201 Created
  private thread · not published

Knowledge compounds when it’s shared. Add your voice to the tree, and together we build it bigger.

Lorediv

Join the beta test, or read the documentation.

Beta access

The Open
Expertise Index

Scattered to
connected expertise

Your work, everywhere

A lore built in public

Open knowledge tree

Discover specialties

Get ahead of the trend

Built to be assembled

Make your content
position you

Be the source

Routing by lore

One address, three formats

Knowledge grows, credit stays

One tree, every format

Layered context

Easy to take part

From readable, to answerable, to reachable

Findable

Queryable

Reachable

Knowledge compounds when it’s shared. Add your voice to the tree, and together we build it bigger.

Lorediv

Privacy & Cookies

What we never do

The only third party

Analytics

Lorediv app is under construction

The OpenExpertise Index

Scattered toconnected expertise

Your work, everywhere

A lore built in public

Open knowledge tree

Discover specialties

Get ahead of the trend

Built to be assembled

Make your contentposition you

Be the source

Routing by lore

One address, three formats

Knowledge grows, credit stays

One tree, every format

Layered context

Easy to take part

From readable, to answerable, to reachable

Findable

Queryable

Reachable

Knowledge compounds when it’s shared. Add your voice to the tree, and together we build it bigger.

Lorediv

The Open
Expertise Index

Scattered to
connected expertise

Make your content
position you