Skip to main content

Drop-in observability and guardrails for AI agents.

Project description

Lightsei

Drop-in observability and guardrails for AI agents.

pip install lightsei
import lightsei
import openai

lightsei.init(api_key="bk_...", agent_name="my-bot")

oai = openai.OpenAI()  # auto-instrumented after init()

@lightsei.track
def reply(prompt: str) -> str:
    return oai.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": prompt}],
    ).choices[0].message.content

That's it. Every call now appears at app.lightsei.com with timestamps, model, latency, and token counts. No instrumentation, no manual wrapping.

What you get

  • Observability — runs, events, costs, errors. Out of the box for OpenAI and Anthropic; one line of code per provider.
  • Guardrails — daily cost caps, output validators (schema + content rules), behavioral checks. Caught before delivery, visible in the dashboard.
  • Polaris — a project orchestrator bot you can deploy via Lightsei's PaaS. Reads your MEMORY.md + TASKS.md and proposes the next moves.
  • Notifications — Slack, Discord, Teams, Mattermost, generic webhook. Polaris's plans land in your team chat, validation failures page you, agent crashes get reported.
  • Graceful degradation, non-negotiable — if Lightsei's backend is unreachable or rejects an event, your bot keeps running. SDK never crashes the user's program.

Configuration

lightsei.init(
    api_key="bk_...",            # your workspace key from app.lightsei.com
    agent_name="my-bot",         # appears in dashboard + cost rollups
    version="0.1.0",             # optional — tags events
    base_url="https://api.lightsei.com",  # default
)

Sign up for a workspace API key at app.lightsei.com/signup.

Deploying bots on Lightsei

lightsei deploy ./my-bot --agent my-bot

Zips the directory, uploads to Lightsei's hosted runtime, builds a venv from requirements.txt, runs bot.py. Logs stream into the dashboard.

Links

Project details


Download files

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

Source Distribution

lightsei-0.1.4.tar.gz (37.5 kB view details)

Uploaded Source

Built Distribution

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

lightsei-0.1.4-py3-none-any.whl (35.3 kB view details)

Uploaded Python 3

File details

Details for the file lightsei-0.1.4.tar.gz.

File metadata

  • Download URL: lightsei-0.1.4.tar.gz
  • Upload date:
  • Size: 37.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.2

File hashes

Hashes for lightsei-0.1.4.tar.gz
Algorithm Hash digest
SHA256 f1604db1cfb7634dc6cb63055a0c3b6ac3f0ae72716b723b1aa9fb0770e43ae1
MD5 d6e68ed7bcc05a8e271b7ba1885206d5
BLAKE2b-256 03c289aceb51f00d208702560774f88aca36635beb3531a01399e900853175eb

See more details on using hashes here.

File details

Details for the file lightsei-0.1.4-py3-none-any.whl.

File metadata

  • Download URL: lightsei-0.1.4-py3-none-any.whl
  • Upload date:
  • Size: 35.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.2

File hashes

Hashes for lightsei-0.1.4-py3-none-any.whl
Algorithm Hash digest
SHA256 8a49db6d39ff667700a2bea241f1e56697e5fcfb2bd16bc8ba358619ddb95d6c
MD5 4c3a38d63e8f3ce9378d069ba9b86806
BLAKE2b-256 c1e83c1d06505466006f7f07ea89b982d3f5dc003034a993ce22c4f38896e2e1

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page