Entries matching “requests”

Holdy

@Ryan_Holdaway

Got Muse logged in to my medical provider’s portal, he pulled the itemized bills, and questioned every line. So far he’s found several times I’d been double billed, asked for some discounts and has saved me over $4,000. If your moat is bureaucracy, you’re cooked.

@ArshansGithub

@ArshansGithub

An OpenAI Responses-compatible shim that serves the Muse assistant as a stateless HTTP model backend for Claude Code via CLIProxyAPI, one model turn per request.

@AshutoshKD

@AshutoshKD

An Express-based Muse connector that books Uber rides from natural language, with routes for fare estimates and ride requests, connector auth middleware and a privacy page.

@raksix

@raksix

A documented setup that runs Muse Spark 1.3 Contributor inside Claude Desktop on macOS through a small Node proxy that translates Anthropic requests to OpenCode Go's OpenAI-style API.

@kamellperry

@kamellperry

A reverse-engineered local CLI and FastAPI wrapper that talks to Meta AI's Muse Spark and exposes an OpenAI-compatible /v1/chat/completions endpoint, with a browser playground and a SQLite request ledger.

@JacobStephens2

@JacobStephens2

A reproducible evaluation of a stochastic defect where Muse Spark 1.1, under some agent-like request envelopes, wrote to claude-smoke.txt when asked for muse-smoke.txt; a neutral control kept 80/80 filenames intact.

You know the trick where you read a DM from the notification shade so it never shows as read? Muse does that for your Instagram DMs, except it can read all of them. I asked it to screen my unread Instagram threads this morning. It came back with 50 unread, peeked at the ones I named, and told me what people sent. Unread count was still 50 afterward. Nothing flipped to read on my end or theirs. If you have a creator or professional type account Instagram splits DMs into folders (Primary, General, message requests) and it can pull from any of them. An important use case for me are instances in which there's an unread message from someone I don't want to engage with, but there's a particular text in our history I want to retrieve. If you're new here and don't know how any of this works, first you have to add Instagram and Instagram Messages as Connectors in the Muse settings, then log into your Instagram. From there, just ask Muse.

A Muse agent receives a one-time secret two ways: the hard way (email link, redacted URL, dropped decryption key — 45 min) and the agent-native way (a Claude agent creates the request with one Qatom call + ~$0.10, the Muse agent retrieves with one free call — 2 seconds, secret delivered once then deleted). Secure Secrets Exchange on Qatom. Tool: https://github.com/StarwaterHeaven/secure-secrets-exchange — full story: https://github.com/StarwaterHeaven/qatom-use-cases/blob/main/stories/sse-credential-handoff-torque.pdf

@Xmarlandoj

@Xmarlandoj

Muse is your personal assistant. Zouroboros Workshop is the workshop behind it. Workshop is an MCP service that runs on a private, single-tenant VPS provisioned for each customer. From a normal Muse conversation, a user can dispatch real software work to a coordinated team of seven AI coding executors — Claude Code, Codex, Gemini, OpenCode, Kimi, Hermes, and Pi — routed by health and capability, with automatic fallback when one executor is unavailable. Three things make it different from a single-vendor hosted builder. First, it runs on the subscriptions the user already pays for: there are no Workshop model credits and no second meter. Second, its memory is shared across every executor and survives across sessions, so a decision recorded by one tool is retrieved by another in a later Muse conversation — memory that compounds instead of resetting. Third, work is verified before it is delivered: builds run through a factory lane with deterministic review, and swarm results return with per-task evidence, not just an agent's claim that it is done. The customer-facing promise is simple: turn your existing AI tools into a workshop that delivers tested, reviewed work — a finished pull request, not a wall of code in chat.

We've been curious how far local models have actually come for agentic coding tasks, so we ran an experiment. Setup: • Model: Muse Glimmer (30B), packaged as a single llamafile • Agent: Hermes coding agent (connected via llamafile's local server mode, zero API keys needed) • Target: Mozilla AI's Otari gateway The Issue: We pointed Hermes at a real, reported bug in Otari (#183) where the gateway returned a vague 502 error on image requests instead of passing through the actual provider error. What the Agent Did: Hermes read the issue, navigated the repo, isolated the bug, created a branch, ran existing tests, wrote a new regression test, and opened a draft PR (#727). All of it ran locally and offline, with zero code written by hand. It's still draft PR territory rather than a merged fix, but it's a solid signal that ~30B local models are getting genuinely capable for real dev workflows, not just toy demos. Video walkthrough of the run: https://youtu.be/5GAgbT-XgHU?si=vJqEDGm9hssCO5-M Happy to answer questions about the setup, model performance, or how Hermes handled tool calling!

Confucius

@ConfuEth

Asked @Muse to file pothole repair requests with the county. Took 30 seconds, zero hassle. Honestly assumed they'd disappear into a void. Five days later: all fixed. 🤯

Cristian Uibar

@cristianuibar

@theo please add support for Muse Code in T3. I've opened a PR here for it using the official SDK. github.com/pingdotgg/t3co…

DHH

@dhh

Got muse cli hooked up for both default agent action and lazy-loading. Come give it a try, @finkd 🙏 github.com/omacom/omarchy…

🚨 AI News | TestingCatalog

@testingcatalog

Since the Meta Muse agent isn't available to everyone, here is a quick UI walkthrough of its web version. Highlights 👀 > The animated Muse avatar is very cool! You can ask Muse to change how it looks, and it will generate the new look along with all the animations. >

@shelajev

@shelajev

A Docker Sandboxes kit that runs Meta's Muse Code CLI in an isolated sandbox, allowing network access only to the Meta services needed for install, auth and model requests.

I've been working on clinical speaker attribution at Omi and wanted to compare the current diarization models on the same audio. I used 15 mock doctor–patient consultations from PriMock57, about 2.4 hours. Full recordings, automatic speaker counts, without telling the models there are two people. Batch Diarization error rate (DER), with ±250 ms boundary tolerance. Lower is better. Model DER Median processing time Pyannote Precision-3 2.891% 18.9 s / recording (API) Nemotron 3 4.803% 0.688 s / recording Pyannote Community-1 6.620% 18.691 s / recording Sortformer v1 6.778% 3.869 s / recording Sortformer v2.1 7.974% 1.077 s / recording VibeVoice-ASR 8.233% 123 s / recording Meta Muse Voice Transcribe † 13.042% 92 s / request (API) Local models ran on one NVIDIA L4. API times include round-trip overhead; VibeVoice-ASR also performs transcription. † Muse used 20 separate clips because of its 10-minute request limit, so its result isn't a whole-recording comparison. Pyannote Precision-3 had the lowest error. Nemotron came next and was the fastest local model. Streaming Model DER Pyannote live API 3.959% Nemotron 3 † 4.971% Sortformer v2.1 † 6.958% VibeVo

Needs a lot of requests compared to Qwen (almost twice) and Gemma (almost x3). Final score is fine, even though it is "not a coding model" https://wonderrico.github.io/local_llm_benchmark/benchmark-main.html more details on https://wonderrico.github.io/local_llm_benchmark/benchmark-detail.html let see Qwen 3.8 tomorrow...

Setup: Home Assistant + Caddy in Docker on a Mac mini at home. Caddy serves only /api/* on the Mac's Tailscale IP, port 8443. Tailscale ACL allows my Muse agent's VM to that port. Goal: let the agent query the HA API remotely. What's verified: • Mac mini is on the tailnet, Tailscale running, peer visible in tailscale status from the agent VM • Caddy serves the API locally (gets expected 401 without token) • Lutron/Ring integrations healthy in HA What fails: every request from the agent VM through Muse's documented tunnel path (HTTP proxy with CONNECT, port 3130) to https://<tailnet-ip>:8443/api/ dies with curl exit 52, "Empty reply from server" — in about 8 seconds, never a proxy error, never a timeout waiting on approval. A stale duplicate node entry in the peer list just times out (exit 28), so that's not it either. This smells like the agent platform's tunnel data plane never passing the traffic, not a Tailscale or HA misconfig — but before I write it off, has anyone here actually gotten a Muse agent VM talking to a tailnet peer? Or found a workaround that doesn't involve exposing things publicly (Funnel/Cloudflare Tunnel/ngrok are the obvious fallbacks)?

"I asked Meta’s Muse to go to YouTube, find a Zuckerberg video about Muse, cut the specific part I wanted, add burned-in subtitles, render it and give me an X-ready video. And it did the whole thing. It searched → found the clip → edited it → added subtitles → rendered → delivered the final file. I didn’t edit anything myself. That’s what makes Muse interesting to me: you’re not just chatting with AI anymore. You’re giving it actual WORK and letting it execute while you do something else. Still very early, but this is a completely different experience." from LorenzoBolsa

adaptlypost.com

adaptlypost.com

AdaptlyPost walks through connecting its scheduler to Muse as a custom connector: give Muse the public OpenAPI spec and docs, then paste a bearer token into Muse's secure credential prompt.

Built a small bridge so Muse can act on my Mac from my phone, and recorded a real session: it looks at the Terminal, reports the epoch, loss and accuracy it sees, then locks the machine when asked. What struck me building it is how much of the work is permissions, not intelligence: per-action consent, small window captures instead of a live feed, rejecting stale observations before any input. Free beta if anyone wants to try it. Developer here, ask away. try wand here today