iamvtor/opencode-setup

Personal Site Search (Exa Python SDK)

Prerequisites

uv pip install exa-py

Set EXA_API_KEY environment variable. Get one at https://dashboard.exa.ai/api-keys

Approach

Use ctx_execute(language: "python", code: "...") to call Exa. Never run Exa searches in main context — always spawn Task agents that return only distilled output.

Supported Parameters

The personal site category supports ALL available parameters:

from exa_py import Exa
import os, json

exa = Exa(os.environ["EXA_API_KEY"])

response = exa.search(
    "building side projects indie hacking journey 2025",
    num_results=15,
    category="personal site",
    type="auto",
    exclude_domains=["medium.com"],  # exclude aggregators for independent blogs
    contents={"highlights": True}
)

results = []
for r in response.results:
    results.append({
        "title": r.title,
        "url": r.url,
        "highlights": getattr(r, 'highlights', [])
    })
print(json.dumps(results, indent=2))

Full Parameter Support

• query (required)
• num_results
• type ("auto", "fast", "instant", "deep-lite", "deep", "deep-reasoning")
• include_domains / exclude_domains
• start_published_date / end_published_date (ISO 8601)
• start_crawl_date / end_crawl_date (ISO 8601)
• include_text / exclude_text (single-item arrays only; multi-item causes 400)
• contents={"text": {"max_characters": N}}
• contents={"highlights": True} / contents={"highlights": {"query": "..."}}
• contents={"summary": True} / contents={"summary": {"query": "..."}}
• user_location
• moderation
• additional_queries
• contents={"livecrawl_timeout": 10000, "max_age_hours": 0} for fresh content