NewAnswers MCP: one question, one checked answer for your agent.See how it works

Company

Jordan Gibbs, founder of Hyperresearch

By , founder of HyperresearchUpdated

Jordan Gibbs founded Hyperresearch and runs it. He wrote the open-source hyperresearch command-line tool, which is the pipeline the hosted service runs, and he edits every guide on this site.

Background

Jordan is a full-stack developer and AI consultant. He has freelanced for years, building software end to end: the backend, the agents inside it, the interface and the deploy. Most of what he has built since 2023 has AI agents at its core.

  • Client builds. Production systems for private clients, including document agents for a professional firm that cite the page each answer came from, a tool that turns legacy code into a specification citing every line, an agent pipeline for a government team that goes from spec to deployed software, and a single agent wired into a small-business owner’s tools and customers.
  • Agent adoption. He sets teams up with Claude Code, Claude Cowork and OpenAI Codex and builds their first workflows on real work, for clients that have included venture firms, legal teams and government. He also works as a fractional chief AI officer for owner-run businesses.
  • Open source and research. Ochre, a free on-device dictation app written in Rust, with text-cleanup models he fine-tuned and published. Research on running many coding agents in parallel, on training language models with ternary weights, and on grounding answers in Wikipedia.
  • Writing. Since 2023 he has written plain-English guides to AI tools on Medium, including beginner’s guides to Claude Code, Claude Cowork and OpenAI Codex.

Why he built Hyperresearch

The deep research tools he used wrote well and cited badly. They attributed quotes to sources that never contained them, cited papers after reading only the abstract, and counted five reprints of one press release as five sources. Hyperresearch is built so that does not ship: every cited claim is checked against the passage it rests on before a report is released.

What he works on

The open-source CLI and the hosted service are the same pipeline, and he maintains both. The design decisions that are his and are visible in the product: the patcher and the polish pass are tool-locked so they can only apply small edits and physically cannot regenerate a draft; the verbatim research question is stored once and re-read by every step, so no fetched page can change the question; the cite-checker and the ship gate receive no configuration from the run’s register or depth settings, so verification never softens by mode; and a retracted citation the document does not acknowledge blocks a report from shipping rather than raising a warning.

Verified

  • He is the author and maintainer of github.com/jordan-gibbs/hyperresearch, published under the MIT licence and on PyPI as hyperresearch. The repository’s commit history, issues and releases are public.
  • He founded and runs Hyperresearch. The company behind it was incorporated in Delaware on 16 March 2026.

Guides he edits

He edits the site’s statistics guides, including consulting rates, small-business valuation multiples and food insecurity statistics, and its guides to buying a business: SBA 7(a) loans to buy a business, business broker fees and buying a franchise. All of them are listed on the guides page.

Elsewhere

Contact

Email support@hyperresearch.ai. That address reaches him for support, privacy requests, security reports and press. Company and entity details are in the Terms; security and subprocessor answers are on trust.