Hacker News

Search the public domain through image embeddings

Hacker News - Tue, 02/10/2026 - 11:09pm

Article URL: https://faenum.com

Comments URL: https://news.ycombinator.com/item?id=46970714

Points: 1

# Comments: 0

Categories: Hacker News

Show HN: Obsidian Visual Skills – Generate Canvas, Excalidraw, Mermaid from Text

Hacker News - Tue, 02/10/2026 - 11:06pm

Hi HN,

I built a set of Claude Code skills that turn text into three types of visual diagrams for Obsidian users.

The Problem:

I take lots of notes in Obsidian. When I want to visualize concepts, I have to manually draw in Excalidraw (time-consuming), write Mermaid syntax (syntax errors, trial and error), or arrange Canvas nodes by hand (tedious positioning).

What It Does:

Three skills, each outputs a different format:

1. Excalidraw Diagram – Hand-drawn style diagrams with CJK font support. Three output modes: Obsidian (.md), Standard (.excalidraw), and Animated. 2. Mermaid Visualizer – Flowcharts, sequence diagrams, state diagrams. Built-in syntax error prevention for common pitfalls (list conflicts, subgraph naming, special characters). 3. Canvas Creator – Obsidian's native .canvas format. MindMap or Freeform layouts with automatic spacing algorithms. Example prompts:

"Create an Excalidraw flowchart showing the CI/CD pipeline" "Visualize this process as a Mermaid diagram" "Turn this article into an Obsidian Canvas" Technical Notes:

Skills are just Markdown files that Claude Code loads on demand. No server setup, no API keys needed (unless you want image export). The Mermaid skill has built-in error prevention rules learned from 50+ common failures.

Repo: https://github.com/axtonliu/axton-obsidian-visual-skills

Demo video and screenshots in the README.

Happy to discuss the error prevention approach or Excalidraw animation mode!

Comments URL: https://news.ycombinator.com/item?id=46970701

Points: 1

# Comments: 0

Categories: Hacker News

AI is now a magic decompiler

Hacker News - Tue, 02/10/2026 - 11:03pm
Categories: Hacker News

AI Can Work on VMs

Hacker News - Tue, 02/10/2026 - 10:59pm
Categories: Hacker News

Medog Hydropower Station

Hacker News - Tue, 02/10/2026 - 10:55pm
Categories: Hacker News

Show HN: Runmoa – Create a global storefront with payments in 1 minute (no-code)

Hacker News - Tue, 02/10/2026 - 10:54pm

Hi HN,

I’m building Runmoa, a no-code platform that lets anyone launch a global storefront in about a minute.

The idea started from watching creators and small businesses struggle to set up simple monetization without developers or complex tools.

With Runmoa, you can create: - a simple brand website - an online shop - live streaming sales - online courses - booking / reservations

all from one place, without writing code.

Payments work immediately after setup (no extra approval steps), and we don’t take a platform fee (only standard PG fees apply).

This is still early, and any sort of feedback will be greatly appreciated: - Is the value proposition clear? - What would stop you from using this? - What’s missing?

Site: https://runmoa.com?from=moa

Happy to answer any questions.

Comments URL: https://news.ycombinator.com/item?id=46970602

Points: 1

# Comments: 0

Categories: Hacker News

Show HN: AdKit MCP: Inject Ads into Your LLMs

Hacker News - Tue, 02/10/2026 - 10:51pm

Article URL: https://www.adkitmcp.com/

Comments URL: https://news.ycombinator.com/item?id=46970573

Points: 1

# Comments: 0

Categories: Hacker News

Ask HN: Is "make me laugh" the next Turing Test?

Hacker News - Tue, 02/10/2026 - 10:44pm

It seems that in the last few years the Turing Test has rapidly fallen off as a useful experiment for testing artificial intelligence. With the right guardrails, the latest frontier LLMs are more than capable of simulating perfectly believable human conversation partners.

Instead, perhaps the next best practical test for separating humans from AIs is essentially, "can it make you laugh?"

The idea here is that LLMs are largely trained to be able to produce statistically likely sentences based on a massive training corpus, and while this allows for incredibly impressive query responses, it is absolutely atrocious at statistically unlikely outcomes, that is, humor.

Humor is often defined by being mentally incongruous which would necessarily make it statistically unlikely - the very opposite of what LLMs are good at.

Thoughts?

Comments URL: https://news.ycombinator.com/item?id=46970533

Points: 1

# Comments: 3

Categories: Hacker News

Show HN: Multi Tenant MCP Platform

Hacker News - Tue, 02/10/2026 - 10:18pm

We couldn’t find much guidance on deploying MCP servers in a multi-tenant setup, so we built one. SageMCP is an open-source platform for hosting isolated MCP server instances with centralized OAuth and API key management.

Each tenant gets its own MCP endpoint (`/api/v1/{tenant}/mcp`) with path-based isolation. Full MCP protocol support — Streamable HTTP, WebSocket, and SSE — with protocol version negotiation, resumable SSE streams, and JSON-RPC batching.

*340 tools across 23 native connectors:*

- *Code & VCS:* GitHub (24), GitLab (22), Bitbucket (19) - *Project management:* Jira (20), Linear (18), Confluence (16) - *Communication:* Slack (11), Discord (15), Microsoft Teams (13) - *Email:* Gmail (14), Outlook (14) - *Documents:* Google Docs (10), Sheets (14), Slides (11), Notion (10), Excel (14), PowerPoint (10), Zoom (12) - *AI coding tool intelligence:* GitHub Copilot (19), Claude Code (18), OpenAI Codex (16), Cursor (18), Windsurf (11) — standardized metrics schema for cross-tool comparison of adoption, cost, and productivity

Plus support for hosting any external MCP server (Python/Node.js/Go) as a stdio subprocess with health checks and auto-restart.

Under the hood: LRU server pooling (5K max instances, 30-min TTL), session management via `Mcp-Session-Id`, per-tenant token-bucket rate limiting, Prometheus metrics, and feature flags for progressive rollout.

Stack: FastAPI, React, SQLAlchemy, PostgreSQL/Supabase, Docker & Kubernetes with Helm charts.

GitHub: https://github.com/sagemcp/sagemcp License: Apache 2.0

Happy to answer questions about the architecture or multi-tenant MCP patterns.

Comments URL: https://news.ycombinator.com/item?id=46970360

Points: 1

# Comments: 0

Categories: Hacker News

Pages