Stack trace for AI
Find what's failing in your AI pipeline and fix it. Plug into your existing observability platform, use our SDK, or call the REST API directly.
Trusted by
Find what's failing in your AI pipeline and fix it. Plug into your existing observability platform, use our SDK, or call the REST API directly.
Trusted by
A user reports an issue. Artanis shows you exactly what the cause was and how to fix it.
Use our SDK or connect your existing observability platform — no code changes required.
# pip install artanis-ai
from artanis import Artanis
artanis = Artanis(api_key="sk_...")
trace = artanis.trace("rag-answer")
trace.input(question="What is AI?", model="gpt-4")
response = llm.generate(prompt)
trace.output(response)// npm install @artanis-ai/sdk
import { Artanis } from "@artanis-ai/sdk";
const artanis = new Artanis({ apiKey: "sk_..." });
const trace = artanis.trace("rag-answer");
trace.input({ question: "What is AI?", model: "gpt-4" });
const response = await llm.generate(prompt);
trace.output(response);# No code changes needed
#
# 1. In Artanis, go to Settings → Integrations
# 2. Click Configure next to Langfuse
# 3. Select your region (EU or US)
# 4. Enter your Langfuse API keys:
# - Public Key (pk-lf-...)
# - Secret Key (sk-lf-...)
#
# Traces sync automatically every 10 minutes.# No code changes needed
#
# 1. In Artanis, go to Settings → Integrations
# 2. Click Configure next to LangSmith
# 3. Enter your API key (lsv2_pt_...)
# 4. Copy the generated webhook URL
# 5. In LangSmith, open your project
# → Automations → Create rule → Webhook action
# → Paste the URL
#
# Traces flow in real-time.# No code changes needed
#
# 1. In Artanis, go to Settings → Integrations
# 2. Click Configure next to Helicone
# 3. Copy your webhook URL and paste it at
# us.helicone.ai/webhooks
# 4. Set sample rate to 100%
# 5. Turn on "Include Enhanced Data"
# 6. Copy the HMAC key back to Artanis
#
# Traces flow automatically.