Try Demo

Record, Inspect, Replay AI Agents

Stop grepping logs. Capture and analyze trace runs for inspection and offline playback.

localhost:4320/session/spacex_investment
provider:vertex status:2xx
#1 IntentSchema200
gemini-2.5-flash1.30s
#2 o3-mini200
o3-mini4.90s
#3 serpapi.com200
Google Search4.60s
#4 UncertaintyDecision200
gemini-2.5-flash2.70s
#5 serpapi.com200
Google Search17.90s
#6 UncertaintyDecision200
gemini-2.5-flash1.60s
Provider
serpapi
Status
200 OK
Latency
4.60s
Tokens
-- / --
{
"search_parameters": {
"engine": "google",
"q": "SpaceX IPO detailed investment analysis and risk factors expert opinions",
"location": "Austin, Texas",
"google_domain": "google.com"
},
"organic_results": [
{
"position": 1,
"title": "SpaceX Share Valuation & Investment Risks - CNBC",
"link": "https://www.cnbc.com/spacex-valuation-risks"
}
]
}
STREAMING ACTIVE

Meet Orchid, the Orchestration interactive debugger

Debug AI, with AI. Orchid gives your autonomous agents native MCP access to untangle pipeline failures instantly, and gives you a rich web UI to audit their work.

Model-Agnostic Proxy

Intercepts standard API libraries (OpenAI, Anthropic, Google Vertex AI) at the transport layer. Zero vendor lock-in, zero manual mock setup.

Model Context Protocol

Exposes tool endpoints directly to AI clients. Search prompt payloads, query capture sessions, export JSON fixtures, or override active target sessions dynamically.

Deterministic Replay

Block live API calls in test suites. The intercepting proxy serves saved local JSON fixtures via pytest decorators (`@replay`), running integration tests offline.

Global Session Overrides

Force capturing or replaying into specific session names via runtime API overrides, SQLite persistence, environment variables, or custom headers.

Debugging AI shouldn't feel like archaeology.

When your LLM is buried deep inside a backend service, finding the root cause of an unexpected output means searching through logs and guessing what was sent over the wire.

Stop sifting through terminal dust.

Orchid captures the exact artifact. Route your traffic through the proxy, and instantly inspect every raw prompt, schema, and response in high-fidelity, without altering a single line of your core business logic.

How it works

client.py
from orchid import OrchidClient
from openai import OpenAI

# 1. Initialize the intercepting client
orchid = OrchidClient(proxy_url="http://localhost:4320")

# 2. Wrap your standard provider library
client = orchid.wrap_openai(OpenAI())

# 3. Traffic is now proxied!
response = client.chat.completions.create(
  model="gpt-5.4",
  messages=[...]
)
Try the Web UI Demo

Licensing & Support

Orchid is a containerized proxy with an open-source SDK.

Local Core

Free

Free to download & run

  • Download the container image
  • Host on local machine, on-prem, or own cloud environment
  • Open source tools (SDK)
  • Local SQLite session recording & playback
Get Local Core

Enterprise

Custom

Hosted service

  • Hosted containers with persistent store
  • Centralized trace storage & sharing
  • Team collaboration features
  • Advanced security, audit logs, & SSO
Contact Us

Ready to debug smarter?

Explore a trace run in our interactive web demo right now.