Roadmap

Sniffr 2026

Sniffr is built by one founder using a custom AI agent infrastructure that ships features at the pace of a small team. Everything below is real, partially built, and on the calendar to land before the end of 2026. We'll move dates as we learn, but not by years.

V1 — Foundation

Live since May 2026

Everything required to run a serious dog-walking business — multi-walker scheduling, intake, payments, reports — built around the dog rather than the invoice.

  • The Window/Horizon scheduling model with a variable per-business commitment radius
  • Self-onboarding QR card → structured intake → auto-Calendar Meet & Greet
  • Voice-to-tagged-events for walks (talk through the walk, get a structured timeline)
  • Photos as durable dog memories with auto-generated hashtags
  • Two-phase auto-positioning: regular-walker inference + statistical time prediction
  • BYO Stripe Connect with billing cycles
  • Multi-service architecture, proven by BarkVentures running in production at Hamilton Bark

V1.5 — Self-Service

Anticipated Q3 2026

Smoothing the walker-acquisition path so onboarding doesn't depend on a setup call, and turning on the AI assists that have been shipping in the background.

  • · Self-serve walker signup at sniffrpack.com → working tenant + Stripe Connect + admin login + branded subdomain end-to-end
  • · AI-assist generate: auto-caption drafts on photos, auto-report drafts on walks (always editable; the walker stays in the driver's seat)
  • · Persisted per-dog personality summaries powering the AI assists
  • · Data import from Time To Pet, Scout, and Precise — making the switch a one-click, not a weekend

V2 — Network

Anticipated Q3 2026

Walkers stop feeling like solo operators. The community starts to act like a labor network — covering each other's days off, sharing trust, getting paid for it cleanly.

  • · Tenant-to-tenant work sharing: hand a Window to another Sniffr walker at their rates
  • · Optional client-side M&G workflow before a handoff (some clients say "I trust anyone you trust"; some don't — the system supports both)
  • · Tenant-to-tenant payout flow via Stripe Connect
  • · Each walker's coverage circle: the trusted nearby Sniffr walkers they can hand work to

V2 is when Sniffr stops being software for one walker's business and starts being connective tissue between businesses.

V3 — Ecosystem

Anticipated Q4 2026

The platform opens — to other AI agents, and to service shapes beyond walks.

  • · Sniffr MCP server: client AI agents like ChatGPT, Claude, and Gemini can read a dog's recent walks, photos, and rolling personality context. When a customer asks their assistant "how was Wally's week?", they get a real answer. (Already built; shipping pending security hardening so a client agent can't be used to bypass tenant rules.)
  • · Multi-service rollout: boarding, daycare, grooming, and training plugins, on the same architecture BarkVentures already proves
  • · Bespoke service shapes: Hamilton Bark designed BarkVenture for mountain trips with their dogs; if you have a service shape that doesn't exist yet, talk to us. Once we build it, every Sniffr walker can use it.

V4 — Community

Anticipated Q4 2026 / Early 2027

The dog-owner side of Sniffr comes online, and the social graph that's been computing quietly in the photo-embedding layer surfaces as Circles and Packs.

  • · Circles — your dog's auto-detected network of dogs they've actually met. Triggered when a recognized friend's owner is at the same park at the same time. Confirmed by the photo uploader. No facial-recognition ML — geofence + app metadata + a one-tap "is that Goldie?" prompt gets us there.
  • · Packs — branded human communities that form around shared parks and times. Invite links, shared photo streams, event management, and a Pack fund.
  • · Walker advertising against Packs: "20% of June's walks booked through this link goes to the Sunday Morning Dachshund Pack."
  • · The referral network made visible, measurable, and economically active.

V4 is when the consumer side of Sniffr starts driving walker acquisition on its own. Owners love their Pack, push their walker to join Sniffr. Walkers join because their Pack is here. The flywheel closes.

How we ship

Most software in this space comes from one of two places. Marketplaces like Rover and Wag — VC-backed, take a cut from both sides of every transaction, and treat your customers as inventory. Or invoicing platforms like Time To Pet, Scout, and Precise — older, slower, originally built around the bill and retrofitted with everything else over the years.

Sniffr is neither. It's a product of Papyrus Labs AI, a research lab working at the intersection of AI and human cognition. (Other Papyrus Labs products include Seshat, structural code intelligence used by AI agents.) Sniffr is the lab's applied dog-care project — and the only one with a working dog-walking business shipping every line of code in production.

Hamilton Bark runs ~130 walks a week, $250K/year in service revenue, on Sniffr today. Every feature listed above survives contact with that business or it doesn't ship.

Your Stripe is your Stripe. Your customers are yours. Sniffr is software, not a marketplace. We don't broker your relationships, take a cut of your customer revenue, or hold your book of business hostage. When you leave us, your customers leave with you. That's not a feature — it's the whole stance, and it's the reason Sniffr exists.

Built fast, by one founder, without outside funding. Sniffr's velocity comes from Papyrus Labs' custom multi-agent build infrastructure — the same class of tooling that lets a small team out-ship a venture-backed eight. That's how a single founder ships an entire year of features in a year, against a real production customer, without taking money that would force the product to extract from the people using it.

Anticipated dates are anticipated, not promised — but the tables are built, the plugins are partially written, and the agents are running. We'll move dates as we learn, but not by years.