sdk-py

Fiber AI Python SDK

Help the user write Python code using the fiberai package.

When to Use

• User wants to BUILD something with Fiber AI in Python
• User says "python", "script", "automation", "data pipeline", "backend"
• User wants a custom integration, ETL pipeline, or automation

Do Not Use When

• User just wants to search or enrich via chat — use /fiber:search or /fiber:enrich instead
• User wants to write TypeScript code — use /fiber:sdk-ts instead

Quick Start

import os
from fiberai import Client
from fiberai.api.search.company_search import sync as company_search_sync
from fiberai.models.company_search_body import CompanySearchBody

client = Client(base_url="https://api.fiber.ai")

result = company_search_sync(
    client=client,
    body=CompanySearchBody(
        api_key=os.environ["FIBER_API_KEY"],
        search_params={},  # Check https://api.fiber.ai/docs/ for filter schema
        page_size=25,
    ),
)

Key Concepts

• The SDK provides typed functions for every API endpoint
• API functions are standalone module functions, not methods on the client — import them from fiberai.api.{domain}.{operation}
• Each module exports 4 variants: sync, sync_detailed, asyncio, asyncio_detailed
• Request bodies are typed model classes (not raw dicts) — import from fiberai.models.{model_name}
• api_key is a field on every body model (for POST) or query parameter (for GET) — not a header
• Search filters go inside search_params (not filters)
• Pagination uses cursor (not page)
• The Client and AuthenticatedClient are the two exported client classes

Schema & Type Discovery

Fiber's request/response schemas are large and evolve across versions. The SDK is the best source of truth:

1. From the installed package: inspect fiberai.models.<ModelName> for every field. Use help(fiberai.models.CompanySearchBody) or IDE autocomplete to discover available fields for search_params, enrichment_type, etc. Model classes enforce correct field names at construction time.
2. From online docs: per-operation pages at https://api.fiber.ai/ai-docs/<operationId>.md describe every field with examples. Start with https://api.fiber.ai/llms.txt for routing.
3. From MCP at runtime: call get_endpoint_details_full("<operationId>") on the Core MCP to get the full current schema.
4. Open-source examples: https://github.com/fiber-ai/open-fiber has working code samples.