DEV Community: Bartosz Bilicki

Hybrid perps in practice: CEX-style books without the usual KYC wall

Bartosz Bilicki — Fri, 31 Jul 2026 00:02:40 +0000

Perp traders usually pick one corner:

CEX: fast books, KYC friction
DEX: self-custody, UX / liquidity tradeoffs

A third lane is hybrid: order-book style execution with a lighter onboarding stance.

Why teams ask for hybrid

International builders hit KYC walls early
Agents need API / MCP access without human spreadsheet ops
Multi-asset books (crypto + some traditional underlyings) beat single-niche DEXs for routines

What we shipped

We run a live hybrid USDT perpetual futures venue:

Product: https://eterna.exchange
App: https://app.eterna.exchange
Tagline we use: CEX performance. DEX freedom. No KYC.

Rough shape:

500+ pairs across crypto / equities / commodities classes
Order book feel for pro charts
Non-custodial angle vs classic CEX custody

Agent path

If you want agents on top of that venue, keep write tools remote and policy-gated:

MCP: https://mcp.eterna.exchange/mcp
Landing: https://ai.eterna.exchange

Open questions for builders

Do you optimize first for latency, custody, or onboarding friction?
For agents, do you prefer venue-native APIs or MCP wrappers?

Curious how others are ranking hybrid vs pure Hyperliquid-style stacks in 2026.

Shipping Telegram Mini Apps in 2026: code when you must, validate first

Bartosz Bilicki — Thu, 30 Jul 2026 23:59:35 +0000

Telegram Mini Apps are just web apps with a Telegram SDK.

That sounds simple until you hit: auth (initData), payments (Stars / TON), distribution, and iteration speed.

What slows teams down

Boilerplate before first user feedback
Payment wiring before the UX is proven
Shipping with zero discovery plan

A practical sequence

1. Validate with a shell
If founders are not engineers, a no-code Mini App builder can produce a live app (UI + backend + Stars/TON) from a plain-language prompt in minutes. We run one at https://apps-father.com (@apps_father_bot).

2. List where people browse
A finished Mini App with no distribution still dies. Public catalogs help. We run https://tg.app for Mini Apps / bots / channels.

3. Then harden
Once retention signals appear, rewrite critical paths in React/Next, add proper initData verification, and keep payments audited.

Payment notes

Telegram Stars are often enough for digital goods inside Telegram
TON fits crypto-native flows
Test refunds and failed payments early

Debugging tip

Prefer a local Telegram WebApp mock / DevTools bridge before phone-only QA. Phone-only debugging wastes days.

Questions

Do you list before or after monetization is live?
What broke first for you: Stars, TON Connect, or initData?

Happy to compare notes with other Mini App builders.

Remote MCP for crypto perps agents: what we learned shipping write-path safety

Bartosz Bilicki — Thu, 30 Jul 2026 23:55:36 +0000

Most "AI trading agent" demos stop at listing markets.

The hard part starts when the agent can place orders.

What broke first

Keys on the model host - local MCP servers with private keys in env look fine in a demo, then become a single-point blast radius.
Write tools always-on - if open/close is available in every chat, one prompt injection is enough.
No sub-accounts - one shared balance for every agent routine makes audits painful.

Pattern that worked better

Keep a remote MCP for execution (HTTP/streamable), so the Cursor/Claude machine never holds exchange keys.
Start read-only (markets, positions, balances).
Gate writes behind policy / human approve.
Prefer sub-accounts per routine or per client.

We shipped that pattern as a USDT perpetual futures MCP:

MCP endpoint: https://mcp.eterna.exchange/mcp
Agent landing: https://ai.eterna.exchange
Underlying hybrid venue (CEX-style books, no KYC): https://eterna.exchange

There is also a Deno-oriented SDK surface (~29 methods) if you want typed clients instead of raw tool calls.

Minimal agent loop

Read markets + positions
Decide size with hard caps
Require confirmation before write tools
Log every tool call

Telegram side note

If your distribution is Telegram-native (Mini Apps / bots), discovery and shipping speed matter as much as the agent:

Catalog: https://tg.app
No-code Mini App builder (TON + Stars): https://apps-father.com

Questions for builders

Do you keep research MCP and execution MCP as separate servers?
What is your default max notional before human approve?

Happy to compare notes.