🔬 Parallel Skill for OpenClaw
High-accuracy web search via Parallel.ai, built for AI agents. Outperforms Perplexity and Exa on research benchmarks.
What it does
- Deep research - cross-referenced facts with citations and excerpts
- Multiple search modes - one-shot (balanced), fast (quick lookups), agentic (multi-hop reasoning)
- Rich results - URLs, titles, relevant excerpts, publish dates
- Company/person research - evidence-based outputs with source links
Quick start
Install the skill
git clone https://github.com/mvanhorn/clawdbot-skill-parallel.git ~/.openclaw/skills/parallelSet up your API key
Get a key from Parallel.ai, then:
export PARALLEL_API_KEY="your-key-here"Example chat usage
- "Use Parallel to research transformer architectures"
- "Deep search for the latest on AI regulation in the EU"
- "Research who's behind Anthropic - founders, funding, board"
- "Fact-check this claim about GPT-5 with sources"
Search modes
| Mode | Use case | Tradeoff |
|---|---|---|
one-shot |
Default, most queries | Balanced accuracy and speed |
fast |
Quick lookups, cost-sensitive | Lower latency, may sacrifice depth |
agentic |
Complex multi-hop research | Higher accuracy, more expensive |
How it works
Uses the Parallel Python SDK (parallel-web). The skill runs a search script that returns structured results with URLs, titles, excerpts, and publish dates. Results include usage stats for cost tracking.
- Docs: https://docs.parallel.ai
- Platform: https://platform.parallel.ai
License
MIT