LR LABS / connect your agent

Compiled cross-border tax law as a deterministic API. Four tools, one endpoint: https://lrlabs.ai/mcp. Same facts + same law = same answer; outside the compiled corridors it refuses rather than guesses. Live observatory →

Claude Code

claude mcp add --transport http lr-labs https://lrlabs.ai/mcp

Claude Desktop / claude.ai

Settings → Connectors → Add custom connector → URL:

https://lrlabs.ai/mcp

ChatGPT (developer mode) / OpenAI Agents SDK

Settings → Connectors → Advanced → Developer mode → Add: https://lrlabs.ai/mcp

# Agents SDK
from agents.mcp import MCPServerStreamableHttp
lr_labs = MCPServerStreamableHttp(params={"url": "https://lrlabs.ai/mcp"})

Cursor

{"mcpServers": {"lr-labs": {"url": "https://lrlabs.ai/mcp"}}}

Gemini / google-genai

# via any MCP client library, e.g.
from mcp.client.streamable_http import streamablehttp_client
async with streamablehttp_client("https://lrlabs.ai/mcp") as (r, w, _):
    ...

Plain REST (no MCP client)

curl -X POST https://lrlabs.ai/api/agent/analyze \
  -H 'Content-Type: application/json' \
  -d '{"treaty":"India-US","facts":{"agentType":"dependent",
       "concludesContracts":"secures","agentExclusivity":"yes"}}'

The tools

tool	answers
analyze_cross_border_tax	PE exposure + Indian liability for a treaty pair — the condition tree, GIVEN set, deciding facts, GAAR gate, authorities
screen_transfer_pricing	safe harbour, TP documentation, method eligibility, arm's-length range (India)
verify_tax_research_note	any tax analysis, checked against compiled law — citations, thresholds, temporal claims
list_compiled_corridors	honest coverage — what this engine can and cannot answer

Discovery: /.well-known/mcp.json · /llms.txt · /openapi.json