#ai-coding-agents
Everything tagged ai-coding-agents — 5 items.
See all tags.
- 08-07 Coding agents: Three new terminal coding agents entered in forty-eight hours
Three separate terminal coding agents launched inside a forty-eight-hour window: Meta's Muse Code in beta, with a contributor tier priced ~12–21x below list in exchange for training data; Memcode in public beta, differentiating on persistent repository memory; and Clark Code, an Apache-2.0 open-source client over a freemium runtime. A fourth entrant in the same window, Mirafold, sells a UI layer over the incumbents rather than an agent of its own — the accessory market arriving.
- 08-06 Mirafold: An MIT-licensed browser UI wraps three terminal coding agents, free unless you go remote
Mirafold launched on Show HN: a locally-run daemon that puts a browser front-end on Claude Code, Codex and Gemini CLI, letting the agent render cards, tables, charts and pinned live widgets alongside text while a real PTY passthrough keeps `sudo` and `ssh` working behind a `!` prefix. The whole product including the relay is MIT-licensed and free to self-host, with a $12/month Pro tier that only buys QR device pairing and an end-to-end encrypted relay for remote access.
- 08-05 Meta: Meta launches Muse Code, pricing a data-sharing "contributor" tier 12x cheaper than standard
Meta launched Muse Code in beta, a terminal coding agent powered by Muse Spark 1.2 that competes directly with Claude Code and OpenAI Codex on macOS and Linux. Standard pay-as-you-go pricing is $1.25/$4.25 per million input/output tokens, but a "contributor" tier undercuts it roughly 12x on input and 21x on output ($0.10/$0.20) in exchange for Meta training on the user's prompts and completions — a discount priced explicitly as a data trade rather than volume or seat pricing.
- 08-05 Memcode: A terminal coding agent bets everything on persistent memory
Memcode launched on Show HN: a terminal coding agent in public beta, shipped as a single static binary with a one-line install, whose differentiator is persistent memory of a repository's history, architecture and decisions across sessions — "every session starts already oriented instead of re-reading everything." Runs on Memcode credits or bring-your-own API keys.
- 07-28 AI devtools: AI devtool ads reach the departure lounge at LaGuardia
A practitioner photographed LaGuardia's departure-lounge screens on 2026-07-27 running back-to-back AI ads: OpenAI's Codex twice, Cognition's Devin once, plus Palantir — out-of-home placement aimed at business travelers, a channel developer marketing almost never uses. One dated observation with photos; no spend data, no campaign attribution beyond the creatives themselves.