Elastik – treating LLM as an HTTP client in less than 200 lines of code

Name: Elastik – treating LLM as an HTTP client in less than 200 lines of code
Availability: InStock
Author: rangersui

by rangersui·Mar 21, 2026·1 point·2 comments

Visit Project View on HN

AI Analysis

●●SolidBig BrainShip It

Treats LLMs like untrustworthy HTTP clients with HMAC audit chains.

Strengths

•HMAC-signed append-only logs provide verifiable audit trails for AI actions.
•Three-mailbox system isolates rendering, execution, and results without complex schemas.
•Protocol fits in ~200 lines, making it auditable and language-agnostic.

Weaknesses

•Reference implementation lacks production features like authentication or rate limiting.
•String-only communication limits complexity compared to structured RPC frameworks like MCP.

Post Description

Elastik is treating LLM as an untrustworthy http client by using mcp as a transparent transportation layer and built the most basic model with a html, database and a server.

I'm using the web security principle to treat this untrustworthy client(LLM) and provide it with the ability every http client can do, so that LLM can write any web app natively without any agentic UI or predefined components, in a safe, physical isolated way.

Similar Projects

AI/ML●Mid

YouDeserveNow, Justifications for Treating Yourself

AI generates fake scientific citations to justify your impulse purchases.

Rabbit HoleCozy

oneprofiledev

301mo ago

AI/ML●Mid

Elia – Governed hybrid architecture (LLM is capability, not authority)

LLM governance framework, but early-stage spec with no working code—Phase 0 skeleton promised.

Big BrainBold Bet

JMC-FR

222mo ago

Developer Tools●●Solid

Orangensaft – A mini Python-like language with LLM eval in lang runtime

Treats model calls as first-class runtime constructs using $..$ blocks with declared-type enforcement, which makes it unusually ergonomic to intermix deterministic code and LLM-powered operations. The Polars-backed dataframe injection that emits structured JSON summaries instead of raw table dumps is a clever, practical touch for token efficiency; neat sandbox for language designers, but explicitly a hobby/toy project rather than something to deploy.

Bold BetWizardry

jargnar

103mo ago

AI/ML●●Solid