Web Search 2, Web Crawl, and Truly Agentic Workflows

Today we’re launching NextDocs v1.1.3 — a major step toward research‑aware, action‑taking AI. Your assistant can now search, crawl, and synthesize from the web, then write directly into your docs and slides with proper citations.

---

Why this matters

Most “search” features stop at pasting links. This release moves beyond that. The assistant now plans a research strategy, gathers evidence across pages, normalizes content, and writes with citations — all while keeping you in control.

Use it to:

• Build market landscapes with sources in minutes
• Summarize a competitor’s docs and product pages
• Compile a brief from 10–20 URLs (reports, news, docs)
• Draft slides that reference real sources you can verify

---

🌐 Web Search 2 — Research That Actually Cites

• Higher‑quality results with better ranking and deduplication
• Inline citations with source titles and domains
• Multi‑pass querying: broad scan → targeted follow‑ups → concise synthesis

The result: faster research with sources you can trust.

---

🕷️ Web Crawl — From One URL to Understanding

• Crawl a domain with depth controls; robots‑aware fetching
• Clean content extraction (no nav, ads, or fluff)
• Automatic chunking and summarization for long pages

Perfect for market scans, competitive intel, or synthesizing reference docs.

---

🤖 Truly Agentic Workflows

The assistant now plans, searches, crawls, and writes end‑to‑end — with clear tool feedback and consistent UI updates. It works across:

• Pageless docs (reports, briefs)
• Paginated docs (multi‑page documents)
• Canvas slides (visual, AI‑editable decks)

---

How it works (under the hood)

1. Plan: The assistant generates a research plan (queries + crawl hints) based on your prompt.
2. Search: It runs multi‑pass queries to cast a wide net, then narrows and ranks.
3. Crawl: For selected domains/URLs, it follows links up to a safe depth.
4. Extract: Pages run through readability filters to remove nav/ads and isolate meaningful text.
5. Chunk: Long pages are chunked, summarized, and attributed to the original URL.
6. Synthesize: Content is synthesized into your doc/slide with inline citations and a source list.

All tool responses are structured, so the UI can show progress and handle edge cases gracefully.

---

What you can build (examples)

• Executive brief: “Give me a 3‑page summary of the top 5 competitors in vertical X with sources.”
• Research deck: “Create 8 slides on LLM eval frameworks with links to docs and comparisons.”
• Policy roundup: “Summarize key AI policy updates in the EU this quarter with citations.”
• Feature review: “Crawl docs.next‑competitor.com/features and summarize relevant pages.”

---

Step‑by‑step: from prompt to finished doc

1. Open a document or slide deck (pageless, paginated, or canvas).
2. Ask: “Research Y. Use Web Search and Crawl where helpful. Create a 2‑page brief.”
3. Watch progress in the tool panel (searching, crawling, summarizing, drafting).
4. Review citations, expand sections, or ask the assistant to refine tone/structure.
5. Export to PDF, Google Slides, PowerPoint, or Word.

---

Quality, speed, and limits

• Results are ranked and de‑duplicated to reduce noise.
• Crawl depth defaults to safe values; adjust depth for focused sites.
• Citations include source title/domain to aid verification.
• Respect robots.txt and content usage norms. Use excerpts; link back to sources.

We’re actively improving multilingual coverage, table extraction, and richer quote handling.

---

✨ Quality & UX Polish

• Cleaner tool messages and progress indicators for research tasks
• Smarter prompts that keep results on‑brief
• Fixes to image fallbacks, gradients, and formatting consistency

---

Release highlights for v1.1.3

• Web Search 2 with citations and better ranking
• Web Crawl with depth controls and readability extraction
• Agentic orchestration with structured tool responses
• Prompt cache refinements for long sessions
• Stability fixes across images/gradients/formatting

---

Try It Now

Ready to create research‑backed content with AI?

Open NextDocs →

New here? Create a free account and start in seconds.

---

What’s Next

We’re expanding agent capabilities with better source clustering, table extraction, and long‑form synthesis controls. If you have a workflow you’d like automated, we’d love to hear it.

📧 hello@nextdocs.io

— The NextDocs Team