lgtmaybe¶

lgtmaybe reviews the code a pull request changes. Pick OpenAI, Claude, Bedrock, Vertex, Azure, ollama, or any OpenAI-compatible endpoint, then run it as a GitHub Action or from your terminal. GitHub gets inline comments and one summary; locally, you get the same findings before you push.

It reads the diff and a little surrounding code, but only comments on changed lines. On GitHub it never checks out or runs the pull request. Generated files and binaries are skipped, secrets are redacted, and all PR text is treated as untrusted.

It checks for:

Correctness and security — logic errors, missed awaits, injection, auth mistakes, and leaked secrets.
Code health — performance problems, needless complexity, deprecations, and risky dependencies.
Tests and documentation — missing or weak tests, stale comments, and undocumented APIs.
Intent — whether the change does what its title, description, and commits say it does.
Ponytail — code you don't need, standard-library opportunities, and simpler ways to get the job done.

Findings are graded from info to critical and land on the exact changed line. A clean change just gets a 👍 LGTM!.

Reviews aren't all it does. /review and /improve run the review, /describe writes a structured overview, /diagram draws a C4-style map of the change, and /ask <question> answers in the PR. Run lgtmaybe diagram to draw the same map locally before you push.

C4Container
    title Async order confirmations after this change
    Person(customer, "Customer")
    Container(web, "Storefront", "React", "Places orders")
    Container(api, "Order API", "Python", "Creates orders (changed)")
    ContainerDb(db, "Order database", "PostgreSQL", "Stores orders")
    Container(queue, "Order events", "SQS", "Buffers OrderCreated events (new)")
    Container(worker, "Notification worker", "Python", "Consumes order events (new)")
    System_Ext(email, "Email provider", "Delivers confirmations")
    Rel(customer, web, "places order")
    Rel(web, api, "POST /orders")
    Rel(api, db, "stores order")
    Rel(api, queue, "publishes OrderCreated (new)")
    Rel(queue, worker, "delivers event (new)")
    Rel(worker, email, "sends confirmation (new)")

Start here¶

Tutorial — Getting started: your first review with ollama, locally and free.
How-to — task recipes: run locally, Bedrock OIDC, Vertex WIF, Azure OpenAI, GitHub Action.
Reference — Configuration: every config field and schema.
Explanation — What gets reviewed, Architecture, Auth model, Data & privacy.

Providers¶

Provider	Auth
`openai`	`OPENAI_API_KEY`
`anthropic`	`ANTHROPIC_API_KEY`
`openrouter`	`OPENROUTER_API_KEY`
`zai`	`ZAI_API_KEY` — GLM / Zhipu AI; optional `--api-base` for the China / coding-plan endpoint
`bedrock`	Ambient AWS creds — GitHub OIDC, no static key
`vertex`	Ambient GCP creds — Workload Identity Federation, no key
`azure`	Ambient Azure AD creds — GitHub OIDC, no static key (or `AZURE_API_KEY`) + endpoint
`ollama`	None — local only, zero cost
`openai-compatible`	`--api-base` to any OpenAI `/v1` endpoint; key optional (placeholder for keyless local servers)

For AI agents¶

The site root publishes a curated llms.txt index of these docs, plus a whole-corpus llms-full.txt, for LLM crawlers and coding agents.