https://www.youtube.com/watch?v=csm6x3carjc

The problem: Most deep research products decide for you when to stop. And since they're subscription-based, the less work they do per query, the more they profit.

What we built: Webhound is a research agent where you set a dollar budget and it keeps searching until it's used. $5 or $100 — every dollar goes into deeper search, more sources, and real verification. The output is either a cited research report or a structured dataset with source URLs per cell.

What's new: Today we're launching Webhound Agent — a full research assistant with a workspace.

• Start research from chat. Describe what you need, the agent proposes a session with an editable budget and model. You confirm before anything runs.
• Multi-step pipelines. Chain steps together — "research X, then extract Y, then write a recommendation" — and they run automatically in sequence.
• Analyze your existing work. Ask the agent about any report or dataset. It reads the full document and answers with citations.
• Organize everything in folders. Nested workspace with drag-and-drop, working directories, and folder-scoped context.
• The agent remembers you. It learns your research interests, budget preferences, and working patterns across conversations.

How it works under the hood: Research sessions run a plan-execute-verify loop. The planner decides what to investigate, the executor searches the web and writes a cited report, and the verifier checks every claim against its sources. When the verifier finds gaps, it sends the work back. This cycle repeats until the budget runs out.

Pricing: Pay-as-you-go. No subscriptions. $5 free credits to start. See real example reports at different budgets: https://webhound.ai/examples