✈️ triply

Trip planning skill for Claude Code and OpenClaw. Tell Claude where you want to go — it pulls live flight prices from multiple airports, digs up Reddit trip reports, finds hotels, checks weather and travel advisories, and builds a day-by-day itinerary. Outputs an HTML brief you can keep.

---

Quickstart (Claude Code)

mkdir -p ~/.claude/skills/trip-planner
curl -o ~/.claude/skills/trip-planner/SKILL.md \
  https://raw.githubusercontent.com/a692570/triply/main/skill/SKILL.md

Then add your profile to ~/.claude/CLAUDE.md once:

TRAVEL PROFILE:
- Home airports: SFO (also check OAK/SJC if cheaper)
- Budget style: mid-range, budget-conscious
- Dietary: needs vegetarian options on menu
- Skip: casinos as primary activity

Then just say: "plan a trip to Cancun, April 14-19"

---

Quickstart (OpenClaw)

mkdir -p ~/clawd/agent-skills/skills/trip-planner
curl -o ~/clawd/agent-skills/skills/trip-planner/SKILL.md \
  https://raw.githubusercontent.com/a692570/triply/main/skill/SKILL.md

Add your profile to your OpenClaw workspace USER.md or memory.

---

What it does

One conversation gives you:

• Flights — prices from all your home airports, flags cheaper alternatives
• Reddit intel — recent trip reports, neighborhood tips, what locals say to skip
• Hotels — 3 options (best value / mid-range / budget), location-aware
• Weather — conditions for your specific dates
• Travel advisory — US State Dept level, flagged if 3+
• Itinerary — day-by-day plan respecting your preferences and skip list
• HTML report — clean, linkable brief saved locally

---

Why a skill vs a CLI

The CLI approach (see triply.py) scrapes approximated data from multiple sources. The skill uses Claude's native research capabilities — actual web search, Reddit, real flight prices via browser — and gives opinionated recommendations instead of raw lists.

For non-technical friends: share this repo and the curl install above. Works with the free Claude.ai interface too if they add the skill content as a custom project instruction.

---

HTML Report Design

The HTML output uses a flat sand/terracotta dark theme (no gradients):

• Background #0f0d0b, cards #161310
• Primary #d97706 amber, accent #c2603a terracotta
• Sections: Flights, Hotels, Weather, Advisory, Activities, Restaurants, Itinerary timeline, Budget breakdown
• All booking links wired (Google Flights, OpenTable, State Dept)

See output/template.html for the template used by the Python CLI.

---

Python CLI (alternative)

Prefer a one-shot command without a conversation:

# Setup
git clone https://github.com/a692570/triply
cd triply
pip install -r requirements.txt
cp .env.example .env  # only needed for Maps + Tavily

# Run
python triply.py --from SFO --to CUN \
  --depart 2026-04-14 --return 2026-04-19 \
  --adults 1 --prefs "outdoor activities, cenotes"

Flag	Description	Default
--from	Origin city or IATA code	required
--to	Destination city or IATA code	required
--depart	Departure date (YYYY-MM-DD)	required
--return	Return date (YYYY-MM-DD)	required
--adults	Number of adults	1
--no-car	Plan around transit/Uber	false
--prefs	Preferences or exclusions	""
--budget	Total budget in USD	none
--html	Export HTML report to path	none

API keys (optional — core features work without them):

• GOOGLE_MAPS_API_KEY — ground transport directions (get one, free $200/mo credit)
• TAVILY_API_KEY — events + restaurants (get one, free tier)

---

Contributing

PRs welcome. Ideas:

• More flight data sources / scraping improvements
• Return flight timing logic
• Booking link integration
• Cached results for repeat queries
• Port to other AI platforms (Gemini CLI, etc.)

---

License

MIT