Haikamera: point your phone at anything and an open-weight model hands you a poem back

Description and use

Haikamera is a mobile web app that instantly generates a short haiku from a photo using an open‑weight vision model, while preserving user privacy.

Problem

Users want a quick, privacy‑respectful way to notice their surroundings by generating a short haiku from a photo.

Outcome

Haikamera delivers a privacy‑first, zero‑dependency mobile web app that generates a short haiku from a photo using an open‑weight vision model, enabling quick, mindful observation without data collection.

Workflow

• Open the Haikamera PWA on a phone or desktop
• Point camera or upload a photo
• Tap "Snap a poem"
• The app downsizes the image to 1024px and re‑encodes it in a canvas to strip EXIF/GPS
• POST the data URL to the server
• Server calls an open‑weight vision model via OpenRouter
• Model returns JSON with title, three lines, objects, colours, mood
• Client parses JSON, displays photo and haiku
• Optionally view journal in localStorage

Evidence

The app's job is to be the shortest possible interruption. There is no feed to refresh and no streak to protect.

Source

The frame is re-encoded in a canvas on your phone, which strips EXIF and GPS before anything is uploaded.

Source

The server speaks plain OpenAI chat-completions. The default model is Google Gemma 4 31B (google/gemma-4-31b-it:free), an open-weight vision model on OpenRouter's free tier.

Source