Hacker News

Article URL: https://arxiv.org/abs/2607.21475

Comments URL: https://news.ycombinator.com/item?id=49055230

Points: 1

# Comments: 0

Whether at work as a software engineer or in my private life, I kept running into the same problem: While working in my terminal, I’d wish I could use AI to make sense of recent output or compose a command. Copying and pasting context between a chat and the terminal was tedious and preemptively running a full CLI coding agent threw away all the efficiency of working in a shell. Both workflows were far from ideal.

What I really wanted was for the coding agents I already used to pop into my terminal when needed, staying invisible the rest of the time.

I couldn't find any existing tools that fit. iTerm2’s AI features don't let me use my existing AI subscriptions, Warp requires changing your entire terminal emulator and workflow (and a new subscription). I wanted a transparent wrapper that could work with my existing tools, staying completely invisible until summoned with a shortcut.

Architecture

I started from the codebase of the excellent mprocs by pvolok to handle terminal emulation and rendering. I used Claude and Codex to strip out the mprocs-specific code, add an on-demand overlay for the AI's terminal, and add an MCP server. Using CLI flags, the MCP server is injected into the AI agent’s config and instructions are added to the system prompt that explain how the AI should interact with the terminal.

The agent can retrieve terminal content, metadata and history, and send input to be approved by the user before being sent to the guest terminal. The AI is also given a CLI command with all the same functionality as the MCP server for agents that don't support MCP or for piping the output instead of dumping it into the AI’s context.

The Hard Part: Making it Truly Transparent

If you are a heavy terminal user, you know how fragile TUI rendering can be. Making the overlay completely seamless required forking the ratatui TUI toolkit and the rat-salsa widget library to solve three low-level problems:

Terminal Scrollback: Most TUIs manage their own buffer without anything going into the terminal’s scrollback, which was a total nonstarter. After several failed attempts to reliably push lines into the scrollback, I landed on using the old-fashioned way: writing them at the top of the screen and appending enough newlines to force a native terminal scroll. This required modifications to ratatui, the TUI framework, to enable it to render this way while maintaining its internal buffer.

Broken Copy-Paste: Empty space at the end of lines was being written to the terminal as literal space characters, which broke native text selection and copying. I fixed this by clearing the rest of the line upon rendering the first empty cell, then continuing from the next line. Every line ending was being copied as a line feed. I fixed that by ensuring that ratatui used the host terminal for line wrapping instead of moving the cursor to the beginning of the next line.

I hope you'll give it a try!

Quick Stats:

License: MIT

Language: Written in Rust (for efficiency, explicit error handling, and letting the type system enforce invariants).

Privacy: Zero telemetry, tracking, or data collection. Logs are stored locally to enable debugging.

Security: The MCP server uses an http server on 127.0.0.1 that is protected by a bearer auth with the secret generated at runtime and passed to the AI agent via environment variable. AI agents that support auth from an environment variable can use it directly and those that don't can use a terminai subcommands as a stdio MCP server which will read the env var and proxy requests to the http server. The CLI tool that offers the same tools as the MCP server operates the same way.

Vibe coded: Unabashedly written mostly by Claude and Codex. I have given a lot of direction to the AI in terms of how to accomplish tasks and demanded extensive testing but most of the code has never been seen by human eyes. The resulting functionality and stability speak for themselves.

Comments URL: https://news.ycombinator.com/item?id=49055227

Points: 1

# Comments: 0

Article URL: https://www.getwren.ai/post/genie-meter-is-on-roi-of-owning-your-genbi

Comments URL: https://news.ycombinator.com/item?id=49055205

Points: 1

# Comments: 0

Article URL: https://news.bloomberglaw.com/ip-law/google-cant-block-ex-npr-host-voice-suit-with-copyright-shield

Comments URL: https://news.ycombinator.com/item?id=49055192

Points: 1

# Comments: 0

Article URL: https://venturebeat.com/infrastructure/microsoft-launches-new-in-house-ai-models-it-says-cut-costs-up-to-89-versus-openai

Comments URL: https://news.ycombinator.com/item?id=49055188

Points: 2

# Comments: 0

Article URL: https://wilburhimself.github.io/blog/64-the-two-sources-of-noise-every-llm-evaluation-system-must-handle/

Comments URL: https://news.ycombinator.com/item?id=49055170

Points: 1

# Comments: 0

Article URL: https://discuss.grapheneos.org/d/40700-grapheneos-protections-against-data-extraction-from-locked-devices

Comments URL: https://news.ycombinator.com/item?id=49055169

Points: 2

# Comments: 0

Article URL: https://chute.codecraftsol.org/

Comments URL: https://news.ycombinator.com/item?id=49055147

Points: 1

# Comments: 1

Article URL: https://shediy.app/

Comments URL: https://news.ycombinator.com/item?id=49055108

Points: 1

# Comments: 0

I'm the founder of Scalattice, this is my second company, third total product. I'm a 2x founder building some challenging software, some easy software, and some curiosity based tools that I've just always wanted to be a part of!

So here is Scalattice.

Scalattice is an OpenAI-compatible inference API. You keep the OpenAI SDK, swap base_url + API key, and call open models (Qwen3, Llama 3.3 70B, Gemma 3, DeepSeek R1, etc.). Inference is performed by a provider node on our distributed network of consumer hosted inference machines, we run an open source agent (https://github.com/scalattice/scalattice-agent) which works in Rust to run the inference and get the provider paid.

In short, anyone can become a provider, offer up their machine with our Windows/Linux Rust agent, and earn some extra cash, or start a farm of machines to make big bucks.

I decided to build some innovative behavioural traits to the API for higher performance/security: 1. output vetting - Scalattice Cloud vets the response against replica responses requested on the API header N times. 2. security tiers - Using some split inference, the job is completed in chunks by multiple providers and in part by Scalattice Cloud hypervisor for added security/privacy (at an additional cost) 3. regional policy - I know how important data residency is to developer clients (from my experience with Digital ID Infrastructure) and so I built the platform to let developers specify the region for inference.

We pay our providers a majority share (Currently 80%) of every token spent by the developer.

We currently back the network with a failover of our own company machines which ensures we never drop a request.

If you want to give it a try in a couple of minutes, we are currently running a "Top Up $10 for UNLIMITED QWEN3" for an entire month of API calling, (terms and conditions apply).

If you want to give it a try in a couple of minutes: 1. Create a key at https://scalattice.cloud/developers 2. Use the copy-paste snippet on the linked post 3. Live rates: https://scalattice.com/pricing

Or, if you want to be a provider (We are really keen to onboard people and get them earning): 1. Create a machine at https://scalattice.cloud/providers 2. Download our Windows app, or use our on-page instructions to curl the agent installation script for Linux. 3. Attach a provider token to the app or Linux agent from the Machine created on the website. 4. Control the machine from the website, select which models you want to offer, and the physical hardware components to use for inference and the hours of operation. 5. Earn!

Happy to take feedback - especially on DX, pricing clarity, and what would make you try this instead of Together / Fireworks / OpenRouter / Salad etc.

This is my first time posting on HN, please let me know if you think the product has potential, there are obviously trade-offs with non-datacentre inference (latency, capability of cards), but I think this could serve a decent amount of developers very well, because not everyone needs H100 cards for their inference.

Thanks again.

Comments URL: https://news.ycombinator.com/item?id=49055093

Points: 1

# Comments: 0

Article URL: https://www.pcgamer.com/games/action/conceived-in-a-secure-military-facility-this-jedi-knight-fansite-has-been-running-consistently-for-almost-30-years-it-looks-really-close-to-what-it-did-back-in-1998/

Comments URL: https://news.ycombinator.com/item?id=49055076

Points: 3

# Comments: 0

Article URL: https://hydra.uvansa.com/

Comments URL: https://news.ycombinator.com/item?id=49055074

Points: 1

# Comments: 0

Article URL: https://phys.org/news/2026-07-syria-ancient-sites-widespread-looting.html

Comments URL: https://news.ycombinator.com/item?id=49055062

Points: 1

# Comments: 0

Article URL: https://www.ams.org/notices/201501/rnoti-p15.pdf

Comments URL: https://news.ycombinator.com/item?id=49054939

Points: 1

# Comments: 0

Article URL: https://notes.hella.cheap/maybe-we-should-revisit-microkernels.html

Comments URL: https://news.ycombinator.com/item?id=49054908

Points: 2

# Comments: 0

Article URL: https://abcnews.com/US/wireStory/americans-rewire-grocery-shopping-routines-digesting-biggest-price-135077702

Comments URL: https://news.ycombinator.com/item?id=49054894

Points: 1

# Comments: 0

Article URL: https://theamericanscholar.org/thanatos-rising/

Comments URL: https://news.ycombinator.com/item?id=49054884

Points: 1

# Comments: 0

Article URL: https://hirejosie.ai/blog/chatgpt-yelp-ai-lead-response/

Comments URL: https://news.ycombinator.com/item?id=49054865

Points: 1

# Comments: 1

So... If PoW (proof-of-work) schemes like Anubis ultimately don't' work in practise, because Bots/Crawlers are increasingly using headless browsers, are able to solve captures, proofs, etc;

What options do we have realistically to filter out or block unwanted traffic?

What clever schemes can we come up that don't rely on centralised serices like Cloudflare?

Comments URL: https://news.ycombinator.com/item?id=49054859

Points: 2

# Comments: 0

Article URL: http://www.warpstock.org/staticpages/index.php?page=ws2026-home

Comments URL: https://news.ycombinator.com/item?id=49054819

Points: 2

# Comments: 0