@spanlens/cli

@spanlens/cli

One-command setup for Spanlens LLM observability.

Validates your Spanlens key against the dashboard, finds out which providers are registered on your project, and AST-rewrites every direct new OpenAI(...) / new Anthropic(...) / new GoogleGenerativeAI(...) in your codebase into the matching @spanlens/sdk factory. One key covers all three providers.

Quick start

npx @spanlens/cli init

That's it. Follow the prompts.

Dry run (see what it would change)

npx @spanlens/cli init --dry-run

Self-hosted Spanlens instance

npx @spanlens/cli init --server-url https://spanlens.yourcompany.com

Points the wizard at your own server: API key validation, dashboard links, and the generated SPANLENS_BASE_URL env var all use your URL instead of spanlens.io.

What it does

🔭  Spanlens setup

  ✓ Detected Next.js (TypeScript)

  Before continuing, make sure you have:
    1. A Spanlens account — https://www.spanlens.io
    2. A Project at https://www.spanlens.io/projects
    3. Provider keys (OpenAI / Anthropic / Gemini) added to that project
    4. A Spanlens key issued for that project (sl_live_…)

  ? Paste your Spanlens key › sl_live_*************

  ✓ Key valid · project chatbot-prod · providers: openai, anthropic, gemini
  ✓ Updated SPANLENS_API_KEY in .env.local
  ✓ Installed @spanlens/sdk (pnpm add @spanlens/sdk)

  ✓ Found 3 patches to apply
    • [openai] app/api/chat/route.ts
        → import: "OpenAI" from 'openai' → { createOpenAI } from '@spanlens/sdk/openai'
        → 1 × new OpenAI(...) → createOpenAI(...)
    • [anthropic] app/api/summary/route.ts
        → import: "Anthropic" from '@anthropic-ai/sdk' → { createAnthropic } from '@spanlens/sdk/anthropic'
        → 1 × new Anthropic(...) → createAnthropic(...)
    • [gemini] app/api/translate/route.ts
        → import: "GoogleGenerativeAI" from '@google/generative-ai' → { createGemini } from '@spanlens/sdk/gemini'
        → 1 × new GoogleGenerativeAI(...) → createGemini(...)

  ? Apply these changes? › yes
  ✓ Patched 3 files
  ✓ TypeScript check passed ✓

  ┌  Next steps  ─────────────────────────────────────┐
  │  1. Add SPANLENS_API_KEY to your deployment env   │
  │     (Vercel / Railway / Fly → Settings)           │
  │  2. Redeploy your app                             │
  │  3. Your requests will show up at:                │
  │       https://www.spanlens.io/requests            │
  └───────────────────────────────────────────────────┘

🎉 Spanlens setup complete

Before / After diffs

OpenAI

- import OpenAI from 'openai'
- const openai = new OpenAI({
-   apiKey: process.env.OPENAI_API_KEY,
-   timeout: 30_000,
- })
+ import { createOpenAI } from '@spanlens/sdk/openai'
+ const openai = createOpenAI({
+   timeout: 30_000,
+ })

Anthropic

- import Anthropic from '@anthropic-ai/sdk'
- const anthropic = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY })
+ import { createAnthropic } from '@spanlens/sdk/anthropic'
+ const anthropic = createAnthropic()

Gemini

- import { GoogleGenerativeAI } from '@google/generative-ai'
- const genAI = new GoogleGenerativeAI(process.env.GEMINI_API_KEY ?? '')
+ import { createGemini } from '@spanlens/sdk/gemini'
+ const genAI = createGemini()

apiKey and baseURL are stripped — every factory reads SPANLENS_API_KEY from env and routes through the Spanlens proxy. Other options (timeout, organization, defaultHeaders, etc.) stay put.

What's supported

• ✅ Next.js (TypeScript + JavaScript)
• ✅ OpenAI / Anthropic / Gemini — auto-detected from your registered provider keys
• ✅ Auto-installs @spanlens/sdk using your package manager (npm / pnpm / yarn / bun)
• ✅ Validates the Spanlens key against the API before writing anything
• ✅ Confirms before overwriting an existing SPANLENS_API_KEY in your env file
• ✅ Runs tsc --noEmit after patching to catch any breakage immediately
• ✅ --dry-run flag (preview without writing or installing)
• ✅ --server-url <url> flag for self-hosted deployments
• ✅ Multiple new XxxClient(...) calls per project
• ✅ Non-destructive env-file writes (preserves comments, other keys)

Coming soon

• Vite / Express / Fastify detection
• Python (FastAPI / Flask) support
• Device OAuth login (no manual API key paste)

Manual integration (if wizard doesn't fit your stack)

See @spanlens/sdk README — the same helpers work without the wizard:

import { createOpenAI }   from '@spanlens/sdk/openai'
import { createAnthropic } from '@spanlens/sdk/anthropic'
import { createGemini }   from '@spanlens/sdk/gemini'

Requirements

• Node.js 18+
• A Spanlens account with a project, at least one provider key, and a Spanlens API key

License

MIT