Skip to main content

shots

Generic, open-source friendly tooling to capture high-res marketing screenshots of a SaaS app.

Key idea: you log in once manually; shots saves a Playwright storage_state.json. After that, it can run repeatedly headless, optionally guided by an LLM (vision) to navigate and/or crop.

Install

python -m venv .venv
source .venv/bin/activate

pip install -e ".[llm,yaml]"
playwright install chromium

Setup

cp shots.yaml.example shots.yaml   # edit with your app's URL and shots
cp .env.example .env                # add your OPENAI_API_KEY

Both shots.yaml and .env are gitignored.

1) One-time manual login

shots login --base-url https://your-app.example.com --out-dir shots_out

This writes shots_out/storage_state.json.

2) Run screenshots from a config

export OPENAI_API_KEY=...
shots run-config --config shots.yaml --use-llm --use-llm-crop --save-source

CLI flags

Flag Default Description
--config (required) Path to YAML/JSON config file
--out-dir from config, or shots_out Output directory (overrides config out_dir)
--use-llm off LLM-driven multi-step navigation
--model gpt-5.2 OpenAI model for navigation/crop
--use-llm-crop off LLM picks a marketing-friendly crop rectangle
--max-crop-retries 2 Crop validation retry attempts
--save-source off Save uncropped source images alongside output
--overwrite off Force re-capture of all shots, ignoring per-shot overwrite settings
--timeout-ms 10000 Page-load / navigation timeout
--action-timeout-ms 5000 Timeout for clicks/typing (fail fast)
--headed off Show the browser window (debug)
--viewport desktop Fallback viewport preset (desktop, laptop, tablet, mobile)
--viewport-w, --viewport-h, --scale from preset Override viewport dimensions
--full-page from config/preset Capture full scrollable page

Config format (YAML)

Simple (flat shots list)

Each shot is auto-wrapped into its own group with output: png.

base_url: https://your-app.example.com
start: /app
out_dir: shots_out

defaults:
  viewport_preset: desktop
  full_page: true
  max_nav_steps: 12

shots:
  - id: dashboard-hero
    description: >
      Capture the main dashboard with KPI cards and a chart visible.
      Close any modal, cookie banner, or tour overlay.
    url: /app/dashboard

  - id: integrations
    description: >
      Show Settings -> Integrations page listing available integrations.
    viewport_preset: laptop

Groups (multi-shot PDFs, labels, folders)

Use groups instead of shots for more control. Each group produces either a single PNG or a multi-page PDF.

base_url: https://your-app.example.com
start: /app
out_dir: shots_out

defaults:
  viewport_preset: desktop
  full_page: true
  max_nav_steps: 12

groups:
  - id: hero-shots
    output: png
    shots:
      - id: dashboard-hero
        description: >
          Capture the main dashboard.
        url: /app/dashboard

  - id: onboarding-deck
    output: pdf
    folder: onboarding
    label: "{id} — {url}"
    label_date: true
    shots:
      - id: step-1-welcome
        description: Show the welcome screen.
        url: /app/onboarding
      - id: step-2-profile
        description: Show the profile setup page.
        url: /app/onboarding/profile

Config reference

Field Level Description
base_url top (required) App base URL
start top Default start path (default: /)
out_dir top Output directory (default: shots_out, overridden by --out-dir)
defaults.viewport_preset top desktop | laptop | tablet | mobile
defaults.full_page top Capture full scrollable page
defaults.max_nav_steps top Max LLM navigation steps per shot (default: 12)
defaults.overwrite top Default overwrite behavior for all shots (default: false)
shots top Flat list of shots (cannot coexist with groups)
groups top List of shot groups (cannot coexist with shots)
groups[].id group (required) Group identifier
groups[].output group png (1 shot max) or pdf (multi-shot)
groups[].folder group Override output subfolder name (defaults to group id)
groups[].label group Label template applied to all shots ({id}, {url}, {title})
groups[].label_date group Append UTC timestamp below label
id shot (required) Shot identifier
description shot (required) What to capture (used as LLM goal)
url shot Start URL for this shot (absolute or relative to base_url)
viewport_preset shot Override viewport for this shot
viewport shot Custom {width, height, scale}
full_page shot Override full-page capture for this shot
label shot Per-shot label override
overwrite shot If false (default), skip when output PNG already exists

Notes

  • --use-llm enables multi-step navigation: the model sees a screenshot + accessibility tree each step and returns one action at a time until it says done.
  • --use-llm-crop asks the model to choose a crop rectangle for marketing-friendly framing, with automatic validation and retry.
  • All navigation is kept same-origin as base_url.
  • A detailed log is written to <out_dir>/shots.log on every run.
  • If the LLM repeats a previously failed action, it is automatically re-queried with a different approach.

Metadata

Release files for shots 0.2.10

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for shots 0.2.10
File Size Uploaded Provenance
shots-0.2.10.tar.gz 36.8 kB With provenance Details

Built distribution (wheel)

Table of built distributions (wheels) for shots 0.2.10
File Interpreter ABI Platform
shots-0.2.10-py3-none-any.whl Python 3 none any Details

Total release size: 73.9 kB

Release files / shots-0.2.10.tar.gz

Download URL shots-0.2.10.tar.gz
Size 36.8 kB
Tags Source
SHA-256 checksum
How to use checksums
5e81624f49b165c8a6ee51295022100b7dcdb3faac2ae3f45a9256c70dc275c2
BLAKE2b-256 checksum
How to use checksums
74881e8e5d061b0c0518470398d334ba366ea439c915190b3b4965f365b0a25a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 25, 2026.

Transparency log

Release files / shots-0.2.10-py3-none-any.whl

Download URL shots-0.2.10-py3-none-any.whl
Size 37.1 kB
Tags Python 3
SHA-256 checksum
How to use checksums
7a04c697ed58e93da6ecba18da48cde35ecc67b1efc097bd77dbae9b812f7694
BLAKE2b-256 checksum
How to use checksums
70af3cedd2ab265c89813a12b80e175e27dfd1c3fc9f60a7ff78e820b7e235f9
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 25, 2026.

Transparency log

Security for shots 0.2.10

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

All 2 files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for shots-0.2.10.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 25, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page