parallel-deep-research

这段内容是审核时保存的快照。外部来源才是完整且最新的版本。

---
name: parallel-deep-research
description: "ONLY use when user explicitly says 'deep research', 'exhaustive', 'comprehensive report', or 'thorough investigation'. Slower and more expensive than parallel-web-search. For normal research/lookup requests, use parallel-web-search instead. Supports multi-turn: pass --previous-interaction-id from a prior research or enrichment to continue with context."
user-invocable: true
argument-hint: <topic>
compatibility: Requires parallel-cli >= 0.3.0 and internet access.
allowed-tools: Bash(parallel-cli:*)
metadata:
  author: parallel
---

# Deep Research

Research topic: $ARGUMENTS

> Requires `parallel-cli` ≥ 0.3.0. If any command below errors with `no such option`, `no such command`, or `unrecognized arguments`, the user is on an older CLI. Tell them to run `parallel-cli update` (or `pipx upgrade parallel-web-tools` if installed via pipx), then retry.

## When to use (vs parallel-web-search)

ONLY use this skill when the user explicitly requests deep/exhaustive research. Deep research is 10-100x slower and more expensive than parallel-web-search. For normal "research X" requests, quick lookups, or fact-checking, use **parallel-web-search** instead.

## Step 1: Start the research

Choose a descriptive filename based on the topic (e.g., `ai-chip-market-2026`, `react-vs-vue-comparison`). Use lowercase with hyphens, no spaces. Reuse this base name in step 2 as `-o "$FILENAME"`.

```bash
parallel-cli research run "$ARGUMENTS" --processor pro-fast --text --no-wait --json
```

The `--text` flag tells the API to return a markdown report (with inline citations) when the task completes, instead of the default structured JSON. Use it for narrative/report-style requests, which is what most users want from "deep research." Drop `--text` if the user explicitly wants structured JSON output.

Optional with `--text`: pass `--text-description "Keep under 1500 words, focus on M&A activity"` to steer length, format, or focus.

If this is a **follow-up** to a previous research or enrichment task where you know the `interaction_id`, add context chaining:

```bash
parallel-cli research run "$ARGUMENTS" --processor lite-fast --text --no-wait --json --previous-interaction-id "$INTERACTION_ID"
```

By chaining `interaction_id` values across requests, each follow-up question automatically has the full context of prior turns — so you can drill deeper without restating what was already researched. Use a lighter processor (`lite-fast` or `base-fast`) for follow-ups since the heavy lifting was done in the initial turn.

This returns instantly. Do NOT omit `--no-wait` — without it the command blocks for minutes and will time out.

Processor options (choose based on user request):

| Processor | Expected latency | Use when |
|-----------|-----------------|----------|
| `lite-fast` | 10–60s | Quick lookups, follow-ups |
| `base-fast` | 15–100s | Simple questions |
| `core-fast` | 1–5 min | Moderate research |
| `pro-fast` | 2–10 min | **Default** — exploratory research, good depth/speed balance |
| `ultra-fast` | 5–25 min | Multi-source deep research (~2× cost) |
| `ultra2x-fast` / `ultra4x-fast` / `ultra8x-fast` | up to 2 hr | Hardest questions, only when explicitly requested |

Notes on the `-fast` suffix: `-fast` tiers use cached web data and are quicker. The non-fast variants (`pro`, `ultra`, etc.) re-fetch fresher data — slower but better for very recent events. Default to `-fast` unless the user specifically asks about news from the last day or two.

Run `parallel-cli research processors` to see the full list with latencies.

Parse the JSON output to extract the `run_id`, `interaction_id`, and monitoring URL. Immediately tell the user:

- Deep research has been kicked off
- The expected latency for the processor tier chosen (from the table above)
- The monitoring URL where they can track progress

Tell them they can background the polling step to continue working while it runs.

## Step 2: Poll for results

```bas