Build on Diogenes

One private intelligence layer for every interface you build

Bring Diogenes models and JAR knowledge into your own products today, with agents and workflows to follow, without creating another disconnected memory system.

The API extends the same intelligence that lives inside the Diogenes workspace.

Read the docs
POST /v1/chat/completions
curl https://api.diogenes.to/v1/chat/completions \
  -H "Authorization: Bearer $DIOGENES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jar_id": "JAR_ID",
    "messages": [
      { "role": "user",
        "content": "What changed since last week?" }
    ]
  }'

API

The same intelligence outside the Diogenes interface

Use the API to bring Diogenes into your own applications, automations, bots, and internal systems.

The knowledge doesn’t have to move into another silo.

Your application can work with the same authorized intelligence underneath it.

One private core

Connect wallet
WORKSPACELive
TELEGRAMComing soon
APILive

Documentation

Start building in minutes

The Diogenes API is OpenAI-compatible, so existing clients and SDKs work with a new base URL and a Diogenes API key.

The docs cover authentication, models, streaming, using JAR knowledge, billing, and examples.

Only the request format is OpenAI's. Nothing is sent to OpenAI.

Existing clients work
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.diogenes.to/v1",
  apiKey: process.env.DIOGENES_API_KEY,
});

const res = await client.chat.completions.create({
  model: "qwen3.5-397b-a17b",
  messages: [{ role: "user", content: "…" }],
});

Scoped access

Give an integration exactly what it needs

An API connection shouldn’t automatically mean access to everything.

A developer can build around a specific JAR, agent, workflow, or capability without exposing the rest of the user’s private environment.

A new key starts with the narrowest useful access: your personal chats only. You add the JARs it needs.

API key · scope
  • Research JARchat
  • Markets JARchat
  • Team JARno access
  • Personal chatsno access
  • Revokestops the key at once

Agent SDK

Coming soon

Build agents around your own workflows

Create specialized agents that operate inside Diogenes’s knowledge and permission system.

A market agent can work from a research JAR. A project agent can maintain a specific workspace. A monitoring agent can watch for change without gaining access to unrelated knowledge.

Specialized agents. Shared private intelligence.

Agents
  • Market agent← research JAR
  • Project agent← one workspace
  • Monitoring agentwatches change only
  • Unrelated knowledgenever granted
Coming soon

Integrations

Coming soon

Connect tools without creating another brain

Diogenes integrations should extend the user’s intelligence layer rather than fragment it.

External apps and services can become part of a workflow while the persistent knowledge remains inside Diogenes.

Connect more tools.

Keep one source of context.

Integrations
  • External apps & servicespart of the workflow
  • Persistent knowledgestays inside Diogenes
  • Memory silos created0
Coming soon

Models

The model is replaceable. The intelligence isn't.

You choose the model, per request or as your account default. Diogenes never switches it for you.

Every model runs against the same knowledge and the same permissions, and states its privacy route. Change the model and your memory and workflows stay exactly where they are.

Diogenes preserves the context around it.

Models
  • Model choiceyours, per request
  • Automatic switchingnone
  • Knowledge & permissionssame for every model
  • Privacy routestated per model
  • Your memory & workflowsunchanged when you switch

Model support

Know exactly what you're calling

Every model lists its profile (fast or deep), context window, tools, credit price per 1M tokens in and out, and its privacy route: private, confidential (attested NEAR AI) or external.

Query the live list from the API instead of hard-coding it.

GET /v1/models
{
  "id": "near-glm-5.3-flash",
  "display_name": "GLM 5.3 Flash",
  "tier": "included",
  "profile": "fast",
  "context_window": 262144,
  "tools": true,
  "credits_per_1m_in": 19.5,
  "credits_per_1m_out": 65,
  "privacy_route": "confidential",
  "confidential": {
    "provider": "near-ai",
    "attestation": "verified"
  }
}

MCP

Take selected Diogenes knowledge into other AI tools

Connect Diogenes to Claude, ChatGPT, Mistral, Cursor or VS Code over MCP, and your private chats, JARs and Consilium open in a panel inside that app. You choose which JARs the connection reaches and how many credits it may spend.

The other app’s model never searches your Knowledge on its own. It sees a Diogenes reply only when you choose Share on it, so you bring the context you choose without exporting your entire private knowledge environment.

MCP · what the app's model sees
  • A reply you choose to shareshared
  • JARs you connectedpanel only
  • Everything elsestays in Diogenes
Claude · ChatGPT · Mistral · Cursor · VS Code

Don't build another stateless AI integration

Build on persistent private intelligence.