AEVION Pro — $19/мес

14 дней бесплатно · Отмена в любой момент · Карта любого банка

AI

QCoreAI

Multi-model AI assistant for the AEVION ecosystem

Why Multi-model AI Engine matters

One AI surface across Claude, GPT-4o, Gemini, DeepSeek, Grok — knows every AEVION module.

Network role: QCoreAI is the intelligence backbone for Multichat, Bank Advisor, Bank Copilot and every future agent. Centralised model usage accounting → predictable per-token economics.

$$$ contribution: Centralised LLM spend across 27 nodes vs. 27 separate API contracts — single biggest OPEX win in the company.

See where this fits in the $1B+ pitch →

🤖

Welcome to QCoreAI

Ask me anything about AEVION — IP protection, signatures, royalties, or how to get started.

QCoreAI capabilities

🧠

Multi-model

Claude, GPT-4, Gemini, DeepSeek, Grok

🔐

Context-aware

Knows your QRight records, Trust Score, history

⚡

Real-time

Instant responses from 5 AI providers

🌍

Multilingual

English, Russian, and 50+ languages

Multi-agent pipeline

Analyst → Writer → Critic with 3 strategies

Regression testing with 6+ judge types

Prompts library

Versioned system prompts, fork & share

Cost tracking, token usage, trends

Budget & Limits

Set monthly spend cap, alert threshold

20 prompts in parallel, one call

Scheduled batches

Hourly, daily, weekly automation

Share session collections with your team

Side-by-side cost + output diff for any two runs

Annotated snippets from run outputs

Configured models, pricing, and health

Full endpoint documentation with request/response

Prompt Playground

Test a prompt against multiple providers at once

Community leaderboard by 👍/👎 feedback score

Pipeline Builder

Design custom multi-step agent chains

Prompt Optimizer

AI-powered prompt improvement suggestions

Agent costs, strategy ratings, peak hours

Starred runs for quick re-access

Personal access tokens for programmatic access

Multi-user teams with shared workspaces

How it works

Three specialised agents collaborate on every run.

1🔬Analyst

Breaks down the prompt, gathers context, and plans the response structure.

2✍️Writer

Drafts the answer based on the analyst output, tuned for depth and clarity.

3⚖️Critic

Reviews, scores, and refines the draft before the final answer is delivered.

Strategies

Pick the agent coordination pattern that fits your use case.

Sequential

Agents hand off in order. Best for: structured reports, technical docs, code reviews.

Parallel drafts

Two writers produce competing drafts; critic picks the best. Best for: creative writing, ideation.

Debate

Pro/Con writers argue; moderator synthesises. Best for: decisions, analysis, balanced views.

270+

API routes

28

Pages

580

Vitest tests

SDK v1.0

npm package

SDK@dosymbek/qcoreai-client · TypeScript / Node / Edge

npm install @dosymbek/qcoreai-client

import { QCoreClient } from "@dosymbek/qcoreai-client";
const client = new QCoreClient("https://api.aevion.app", { token: process.env.QCORE_TOKEN });
const run = await client.runMultiAgent({ input: "Summarise Q1 2025", strategy: "sequential" });
console.log(run.finalContent);

Ready to build?

Start with the multi-agent pipeline — Analyst, Writer, Critic in one call.

Start building →