Peerd
Peerd is a browser-native AI agent harness that turns Chrome or Firefox into a local agent workstation with your tabs, sessions, and model
Peerd bets where most agent projects won't: the browser is already your best sandbox, session store, and runtime, so build inside it rather than beside it. For a developer who wants an agent inheriting real logins without OAuth gymnastics or a headless browser farm, the shape is compelling — a real Debian VM in a tab, notebooks in sealed workers, P2P agent networking over WebRTC. The tradeoffs are maturity and audience: load-unpacked install from GitHub today, and not much for someone who wants
Verified 5d ago · liveness 73/100 · cite: rightaichoice.com/tools/peerd
- Developers who want a local, private agent controlling their real browser with true session state.
- Power users automating web tasks without data or model keys leaving their machine.
- Privacy-conscious tinkerers avoiding cloud-based agent services and headless browser farms.
- Researchers exploring browser-native agent architecture and peer-to-peer agent coordination.
- Non-technical users who want a plug-and-play AI assistant with no setup.
- Teams needing multi-user collaboration or cloud-hosted agent access.
- Enterprises requiring centralized management, audit trails, or compliance controls.
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip peerd if you want a one-click Chrome Web Store or Firefox AMO install and a plug-and-play assistant — peerd is pre-release, installs load-unpacked from GitHub, and expects you to bring your own provider key or run Ollama locally.
Peerd itself is free and open source, but you pay your model provider directly — an Anthropic or OpenRouter key is billed per token by that vendor, and heavy agent use adds up without any peerd-side cap.
Peerd is free and open source (Apache 2.0) with no peerd-side subscription for any stage — solo developers and small teams both run it at $0 software cost. Your real cost is the model provider you attach: a paid Anthropic or OpenRouter key billed per token, or free local inference via Ollama if your GPU can carry it. Compared with cloud browser-automation services that meter per session, peerd pushes that spend to whichever provider you already pay.
In short
Peerd — Peerd is a browser-native AI agent harness that turns Chrome or Firefox into a local agent workstation with your tabs, sessions, and model. Best for Developers who want a local, private agent controlling their real browser with true session state., Power users automating web tasks without data or model keys leaving their machine., Privacy-conscious tinkerers avoiding cloud-based agent services and headless browser farms.. Free to use.
What's new in Peerd
Checked 5 days agoAcross the latest 2 updates: 1 launch and 1 news mention.
Mixture of Actors
peerd describes its 'mixture of actors' harness design: multiple specialized loops instead of one monolithic agent loop, isolated by the browser sandbox.
Introducing peerd
peerd launched publicly via Show HN: an Apache 2.0 browser extension that runs an AI agent harness client-side, with local or remote model keys.
What people actually say about Peerd — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
36 mentions across 2 sources (Hacker News, Lemmy) · researched Jul 3, 2026.
Average across the 2 sources that answered — each source counts once, not each post.
- +Runs entirely in the browser with no background processes.
- +No telemetry; user's data never leaves their machine.
- +Supports BYOK and local models for privacy.
- +Full access to browser tabs, sessions, and extensions.
- +Sandboxed compute with JS notebooks and WASM VMs.
- −Documentation is verbose and reads like AI slop.
- −Extension installation is more friction than a web page.
- −Feature list is comically extensive and overwhelming.
- −Project name confusion with Microsoft's Peerd fork.
- −Limited community beyond Hacker News.
- • Requires own API keys for cloud models; no free model included
Viability Score
How well maintained and how widely used is Peerd? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Open-source browser extension for Chrome and Firefox (Apache 2.0)
- Mediated browser tools that drive your real tabs, sessions, and web apps
- Bring your own model key for Anthropic and OpenRouter
- Keyless local model inference via Ollama at localhost:11434 or an on-device WebGPU runner
- Local voice transcription with Moonshine over WebGPU — audio stays in the tab
- JS notebooks running in a sealed Web Worker with OPFS storage
- Compiled WASI tools (SQLite, codecs, language runtimes) in sealed workers
- Real Debian Linux VMs compiled to WebAssembly (CheerpX), each in its own tab
- Agent memory, planning, subagents, and budget/permission limits
- Plan mode by default; Act mode required before clicking, typing, or submitting
- Peer-to-peer agent-to-agent coordination over an N-peer WebRTC mesh
- Ed25519 identity (did:key) and Kademlia DHT discovery
- Content-addressed peerd:// bundles for sharing apps and workflows
- WebCrypto AES-GCM vault with passkey-first unlock and 45-minute idle auto-lock
- Five modules: Act, Think, Compute, Build, and Share
About Peerd
Peerd is an Apache 2.0 browser extension that runs a full AI agent harness client-side inside the Chrome or Firefox you already use. Rather than a new AI browser you migrate to or a cloud automation service driving a headless instance, it drops the agent into your existing tabs, logged-in sessions, and extensions. There is no Bun server, Go binary, Python sidecar, Docker container, or companion process. Workload execution stays in the browser. Peerd wraps the V8 isolate into Linux VMs via WebAssembly (CheerpX) and sealed Web Worker notebooks backed by OPFS, each running in its own tab, with sandboxed iframe apps. The harness organizes around five primitives: Act (mediated browser tools that drive real tabs, sessions, and web apps), Think (model routing, subagents, memory, budgets, permissions), Compute (notebooks, JS workers, compiled WASI tools, and Debian VMs), Build (local artifacts, apps, workflows), and Share (peer-to-peer WebRTC agent networking). Modern touches matter here too. Bring your own Anthropic or OpenRouter key, or run keyless inference against Ollama at localhost:11434. Voice is off by default; when enabled, local transcription via Moonshine runs over WebGPU so audio stays in the tab. A WebCrypto AES-GCM vault with passkey-first unlock, auto-lock after 45 minutes idle, an Ed25519 did:key identity, content-addressed peerd:// bundles, and a Kademlia DHT round out the privacy story. Launched via Show HN in June 2026, peerd describes its architecture as a 'mixture of actors' — multiple specialized loops rather than a single monolithic agent loop, isolated by the browser sandbox. Peerd is aimed at developers and privacy-focused tinkerers comfortable loading an unpacked extension. It is not a plug-and-play assistant like the chat apps, and it is not a cloud operator like Browserbase. Pick it if you want your model keys and session state to stay on your machine and you are willing to install from GitHub while store listings are pending.
Behind the Verdict
I keep recommending Peerd to one specific person: the developer who already lives in twenty logged-in tabs and resents piping their cookies into a cloud agent. The premise is honest — the browser is the runtime, the network, the hypervisor, and the operating system, and Peerd just stops pretending otherwise. You install an extension; the agent gets your tabs, your DOM, your sessions; your model keys never leave the machine. That's a real answer to a real problem. What makes it more than a gimmick is the execution layer underneath. A genuine Debian VM compiled to WebAssembly running in a tab means the agent gets POSIX without a remote shell. JS notebooks in sealed Web Workers with OPFS, compiled WASI tools like SQLite, and a WebCrypto AES-GCM vault with passkey unlock — none of that is marketing. The 'mixture of actors' design it wrote up in July 2026 also tracks how serious agent builders are moving: multiple specialized loops, isolated by the sandbox, instead of one omnivorous agent. Where it bites is setup. You load it unpacked from GitHub. Chrome Web Store and Firefox AMO are listed as coming soon, so day-to-day friction is real. If you want a one-click install and someone else to handle updates, this is not there yet. Pass on Peerd if you need multi-user collaboration, cloud-hosted agent access, or the audit trails and compliance controls an enterprise security team will ask for. There's no centralized management console and there isn't going to be one soon — it is deliberately a single-user, local-first tool. The closest alternative depends on what you're optimizing for. If you want cloud reliability and a headless fleet, Browserbase or Playwright MCP are the obvious picks, with the obvious tradeoff that your session state goes with them. If you want an
Researching Peerd? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Peerd actually fits — and what changes day-one when you adopt it.
You open the peerd side panel in the browser you are already signed into, create a session, and describe the form-filling task. The agent starts in Plan mode, reads the page via the accessibility tree and your real cookies, then you switch to Act so it can click, type, and submit.
Outcome: The task completes in your real session without OAuth setup, credential storage, or a headless browser farm, and side effects pause for sign-off until you disable the confirm toggle.
You load a CSV into a peerd notebook running in a sealed Web Worker with OPFS, or spin up a Debian Linux VM in its own tab, and have the agent write and run the analysis without any cloud container.
Outcome: The data never leaves your machine — execution stays in the browser's V8 isolate, and you can watch, kill, or attach to the VM tab from any chat session.
You run peerd in two browsers, let them discover each other over the Kademlia DHT, and exchange capabilities across the N-peer WebRTC mesh using Ed25519 did:key identities.
Outcome: You get a working A2A prototype where the only servers are stateless introducers that drop out of the data path, and you can share signed apps or workflows as content-addressed peerd:// bundles.
Use Cases
- Automate web workflows using your own browser session and logged-in accounts, like filling forms or browsing on your behalf.
- Run local data analysis by spawning JS notebooks or Debian Linux VMs in a tab, keeping sensitive data off cloud servers.
- Build a personal AI assistant that acts on real web pages privately, with nothing leaving your machine.
- Prototype agent-to-agent communication over WebRTC between two peerd instances for distributed tasks.
- Use local Moonshine-over-WebGPU transcription for voice input without sending audio to the cloud.
- Run a throwaway Linux VM in a tab to test destructive commands where the worst case is a broken VM, not your home directory.
Models Under the Hood
as of 2026-10-18
Limitations
- Peerd is pre-release: Chrome Web Store and Firefox AMO listings are still 'coming soon', so installation is from GitHub.
- It lives entirely inside your existing desktop browser runtime as an extension, with no background process, sidecar, or cloud server, and you bring your own model keys (Anthropic, OpenRouter, or keyless Ollama/WebGPU local inference).
- It assumes familiarity with developer-mode extension installs, WebAssembly, WebRTC, and browser tooling.
as of 2026-10-03
Verification history
We have re-verified Peerd 9 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 9 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Peerd tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Open Source
$0
Ideal for
Developers, power users, and privacy-focused tinkerers who are comfortable installing load-unpacked from GitHub and supplying their own Anthropic, OpenRouter, or local Ollama model access.
What this tier adds
Starting tier: the full extension at $0 under Apache 2.0 — no account, no backend, no telemetry, and no paid tier above it.
Where the pricing makes sense
The company stage and team size where Peerd's pricing actually pencils out — and where peers do it cheaper.
Peerd is free and open source (Apache 2.0) with no peerd-side subscription for any stage — solo developers and small teams both run it at $0 software cost. Your real cost is the model provider you attach: a paid Anthropic or OpenRouter key billed per token, or free local inference via Ollama if your GPU can carry it. Compared with cloud browser-automation services that meter per session, peerd pushes that spend to whichever provider you already pay.
Setup time & first value
How long it actually takes to get something useful out of Peerd — broken out by persona, not the marketing-page minute.
For a developer: roughly 15-30 minutes end to end — clone the repo, enable Developer Mode, Load Unpacked the extension/ directory (no build step), then enroll a WebAuthn passkey on first run to create the vault. From there, pasting an Anthropic or OpenRouter key in Options → Providers takes a minute, and pointing at local Ollama skips keys entirely. Non-technical users should not expect a quick
Switching to or from Peerd
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From a terminal coding agent (Claude Code, OpenCode, Aider): keep it for file edits, and add peerd when the task needs the page on your screen, a form filled, or the session you are logged into.
- →From a cloud browser-automation service (Playwright MCP, Browserbase): move tasks that depend on your real logged-in state into peerd so the agent and page share one process on your machine.
- →From a separate AI browser (Comet, Atlas, Dia): keep the browser you already use and add the agent as an extension, preserving your tabs, sessions, and extensions.
- →From a cloud model gateway: paste your Anthropic or OpenRouter key into the peerd vault, or run Ollama locally at localhost:11434 keyless.
- ↗To Claude Code or Aider: hand off tasks that need file-level editing or terminal execution in a repo, since peerd cannot see your filesystem.
- ↗To Browserbase or Playwright MCP: move high-volume headless automation that does not need your personal session cookies onto a managed cloud runner.
- ↗To a plug-and-play chat assistant: switch if you decide load-unpacked installation and BYOK are more setup than you want.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Peerd”, and we withheld 6: 6 could not be judged, because “Peerd” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Peerd.
Official links
Tools that pair well with Peerd
Common stack mates teams adopt alongside Peerd, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Peerd vs Spider Cloud
Peerd is ideal if you need a private, local agent that controls your browser and keeps data on your machine — perfect for tinkering with agent architectures and privacy-sensitive automation. Spider Cloud is the better choice for web data extraction at scale, especially for powering RAG pipelines and AI agents that rely on up-to-date content. Choose Peerd for client-side control and privacy; choose Spider Cloud for production-grade crawling and cloud integration.
Peerd vs Presto Voice
Peerd and Presto Voice serve completely different markets. Peerd is a free, open-source browser extension for developers who want a private, local AI agent to automate browser tasks. Presto Voice is a contact-priced enterprise platform for QSR chains to automate drive-thru orders and boost revenue. Your choice depends on whether you need a tinkerer’s agent toolkit or a turnkey voice AI for restaurants.
Peerd vs Temporal Ai
Peerd is ideal for individual developers seeking a private, local AI agent that controls the browser directly, with no cloud dependency. Temporal serves teams needing enterprise-grade durable execution for AI agents and workflows, with automatic retries and persistence. Choose Peerd for personal agent automation that stays on your machine; choose Temporal for production-grade orchestration that survives failures.
Alternatives to Peerd
View all
Popular in Browser & Computer-Use Agents
Spider Cloud
Spider Cloud is a web scraping, crawling, and search API that streams live pages as markdown or JSON for AI agents and RAG pipelines.
OpenAgents
OpenAgents is an open-source language agent platform for data analysis, 200+ plugins and autonomous web browsing.
Frequently Asked Questions
Best-of guides
Used Peerd? Help shape our editorial sentiment research.