Skip to main content

Agentic Flights

Give an AI your trip plans. It searches Google Flights across airports, dates, and trip lengths, compares the best options, and checks current prices before you choose.

Give this to your AI

Copy this prompt into Codex or Claude Code and replace the trip details:

Install Agentic Flights with:
`python -m pip install -U agentic-flights`

Install its reusable skill for the AI app you are using:
- Codex: `agentic-flights init-skill --scope user --harness codex`
- Claude Code: `agentic-flights init-skill --scope user --harness claude`

Use the installed Agentic Flights skill to find the best flights for my trip.
Explore flexible dates, nearby airports, and trip lengths when they
could improve the result. Show options that balance price and travel time. Check
the current total price and baggage rules for the best options. Tell me if any part
of the search was incomplete. Do not book anything.

Trip: [origin, destination, dates or flexibility, trip length, travelers,
currency, baggage, stop limits, and timing or airline preferences].

Install it yourself

Agentic Flights requires Python 3.11 or newer. Discovery uses direct HTTP requests. Final-price verification uses a tool-managed browser. Install Chromium for it:

python -m pip install -U agentic-flights
python -m playwright install chromium

Install the skill in the current project for both Codex and Claude Code:

agentic-flights init-skill

This adds the skill to .agents/skills for Codex and .claude/skills for Claude Code. Add --scope user to use it in every project. It will not replace a changed skill unless you pass --force.

What it can do

  • Search one-way, round, and multi-city trips, including trips that return from a different city or to a different airport.
  • Explore nearby airports, dates, and stay lengths; filter by price, stops, airline, connection, layover, time, duration, emissions, and baggage.
  • Search for adults, children, and infants, including lap infants.
  • Compare price against travel time, departure hour, and time at the destination without mixing different currencies.
  • Check the current total price, connection-level same-flight match, baggage rules, and booking links.
  • Say when it could not read a flight, stopped early, or still has work to do.

Ways agents can use it

The skill is the easiest way to start. Broad searches use a fast direct request; final-price and baggage checks use a background browser. The package also has a Python API, a command-line interface, and an optional local MCP connection for AI apps:

python -m pip install -U "agentic-flights[mcp]"
agentic-flights-mcp

Large CLI searches return resumable work chunks. Progress and checkpoint run IDs appear on stderr; the final JSON includes batch_progress and a resume_command. Completed queries survive interruption. Run agentic-flights resume RUN_ID to continue.

Agents can pass an exact search directly with AgentAPI.start({...}); no temporary request file is needed. Discovery stays HTTP-only. A headless browser starts only when final prices or baggage are verified.

Run agentic-flights guide for the workflow or agentic-flights guide <operation> for one operation. Saved searches stay on your computer and expire after seven days without use by default. The project does not run a hosted search service.

Know before relying on a result

This unofficial project is not affiliated with Google. Google Flights has no supported public search API, so page changes can break the tool. Results show what was visible at search time; they are not reservations or price guarantees. The tool cannot prove that Google showed every fare. Confirm the price, baggage, and seats with the seller. The tool does not book travel.

Agent guide · Technical reference · PyPI · Contributing · Security · Third-party notices

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

agentic_flights-0.6.7.tar.gz (170.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

agentic_flights-0.6.7-py3-none-any.whl (81.2 kB view details)

Uploaded Python 3

File details

Details for the file agentic_flights-0.6.7.tar.gz.

File metadata

  • Download URL: agentic_flights-0.6.7.tar.gz
  • Upload date:
  • Size: 170.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.6.4

File hashes

Hashes for agentic_flights-0.6.7.tar.gz
Algorithm Hash digest
SHA256 b31713cc2966077ee9a508dbe572a740f96a1c286ee44fe539c91d126b659480
MD5 e97baf3ab1d096261384cf44dff08250
BLAKE2b-256 06f908540dfb58614cb38da957352ad6a0cd9b1c1765cb6f583fbf845e23fe58

See more details on using hashes here.

File details

Details for the file agentic_flights-0.6.7-py3-none-any.whl.

File metadata

File hashes

Hashes for agentic_flights-0.6.7-py3-none-any.whl
Algorithm Hash digest
SHA256 1fe42923ebdffe56f9aadada09c6179be8f054d3fc33ba9a7e465f14d2969f21
MD5 158dd1c48f42463a56e9c7692aef325b
BLAKE2b-256 66c2f843aecd9d5ee26306651efd0b7aab30446cfff68e78507ada8d9c0c5380

See more details on using hashes here.

Release history Release notifications | RSS feed

0.7.0

2 files

0.6.8

2 files

This release

0.6.7 This release

2 files

0.6.1

2 files

0.6.0

2 files

0.5.0

2 files

0.4.4

2 files

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