OpenAI 2026 hackathon

Second Witness

Second Witness is an evidence intelligence workspace that helps investigation teams convert fragmented source material into transparent, stress-tested and auditable cases.

Hackathon project · 0 likes · 0 comments

Archive position — measured, not model output

0 likes on Devpost

2,264 of the 7,856 archived projects have more likes, and 5,592 share exactly 0 — so this project's #6,603 place in the like-ranked listing is a tie-break inside that group, not a ranking.

Projects (log scale)

1
10
100
1k
10k
05,592
11,758
2285
3–4132
5–975
10+14

Likes on Devpost. ▲ marks this project's group.

Show the figures
LikesProjectsShare of archive
05,59271.2%
11,75822.4%
22853.6%
3–41321.7%
5–9751.0%
10+140.2%
Devpost like counts for all 7,856 archived projects, captured when this archive was built.

Executive Summary

What the company appears to be

Second Witness is an AI-powered evidence intelligence workspace designed for investigation teams. The author states it helps transform fragmented source material into transparent, stress-tested and auditable cases. It uses AI agents to extract claims, build timelines, detect contradictions, and support adversarial review workflows.

What changed

The project description indicates a self-developed tool built over time by one person (the author), with no team or funding mentioned. It was submitted as part of the OpenAI 2026 hackathon, suggesting it is in early development or prototype stage.

Single most important open question

Is there any evidence that this product has been used in real-world investigations or tested by actual users? The description contains no mention of customers, adoption, revenue, or usage metrics beyond the author’s personal experience and vision.

Back to contents

What The Product Actually Is

The description states:

  • Second Witness is an AI-powered evidence intelligence workspace.
  • It allows users to upload various forms of source material (documents, screenshots, audio, emails, spreadsheets, links).
  • It transforms this into an interactive investigation workspace with features like:
    • Claim extraction
    • Evidence graph
    • Investigation timeline
    • Contradiction detection
    • Source independence analysis
    • Stress-Test This Case workflow using AI agents (supporting researcher, contradiction hunter, source auditor, alternative explanation agent, case editor)
    • Evidence-backed follow-up questions
    • Privacy protection and redaction tools
    • Auditable investigation reports

Inference: The product is a full-stack application built with modern web technologies including React, Node.js, OpenAI APIs (GPT-5.6), and others.

Back to contents

Positioning & Claim Evolution

The description states:

  • The tool is not an automated truth detector, but rather a system that shows what is claimed, where it came from, which evidence supports or challenges it, and what remains unresolved.
  • It aims to reduce confirmation bias by enabling adversarial workflows.
  • The name “Second Witness” is inspired by the idea of testimony being examined, supported, and verified rather than accepted without evidence.

Inference: Positioning is centered on evidence transparency, human control, and reducing bias in investigative processes. It positions itself as a tool for journalists, researchers, and investigation teams seeking structured, auditable outcomes.

Back to contents

Target Customer & ICP

The description states:

  • The primary users are journalists, researchers, and investigation teams.
  • It is intended to help those working with scattered source material in complex investigations.
  • The author personally chose the name inspired by the need to support his brother's work as a journalist.

Inference: The ICP likely includes individuals or small groups involved in investigative journalism, fact-checking, legal research, and compliance roles who deal with unstructured data and require tools for organizing claims and evidence.

Back to contents

Business Model & Pricing Evidence

The description states:

  • No explicit business model or pricing information is provided.
  • There is no mention of monetization strategy, subscription plans, or commercial use cases beyond the author’s personal motivation.

Not evidenced: No indication of how the product will be sold or whether it will be offered as a SaaS platform, open-source tool, or other format.

Back to contents

Technical & Delivery Signals

The description states:

  • Built with React, Node.js, OpenAI API (GPT-5.6), Codex, and various cloud services.
  • Uses structured AI outputs instead of uncontrolled text generation.
  • Implements multi-agent workflows with clearly separated responsibilities.
  • Includes features like evidence graph, document viewer, timeline reconstruction, contradiction detection, etc.
  • Designed for privacy protection including redaction of sensitive data.

Inference: The technical stack suggests a modern full-stack web application with AI integration. The use of structured outputs and agent orchestration indicates an emphasis on traceability and control over AI behavior.

Back to contents

Traction & Maturity Signals

The description states:

  • Team size: 0
  • No mention of customers, users, or adoption metrics.
  • No revenue, ARR, headcount, or funding rounds are mentioned.
  • The project was submitted to a hackathon (OpenAI 2026), indicating early-stage development.

Not evidenced: There is no evidence of traction, user base, or product maturity beyond the author’s own development efforts.

Back to contents

Competitive Context

The description states:

  • No direct competitors are named.
  • It is positioned as a tool for investigative teams, not general-purpose AI tools or document management systems.
  • The focus on adversarial workflows and stress-testing suggests it may differentiate from standard AI summarizers or knowledge bases.

Inference: Likely operates in a niche space related to investigative research, fact-checking, and evidence-based decision-making. May compete with tools like Notion, Airtable, or specialized legal research platforms, though not clearly defined.

Back to contents

Key Risks & Red Flags

The description states:

  • The author is the sole developer.
  • No team, funding, or commercial traction are evident.
  • The product is described as a personal project, not a scalable business.
  • It relies heavily on AI (GPT-5.6) and may be vulnerable to hallucinations or misinterpretations if not carefully validated.
  • There is no mention of data privacy compliance, scalability, or long-term sustainability.

Inference: Key risks include:

  • Lack of team or funding
  • Unclear path to market or monetization
  • Heavy reliance on AI outputs without clear validation mechanisms
  • No evidence of real-world testing or user feedback

Back to contents

Diligence Questions To Ask The Founders

  1. Has the tool been tested with actual investigation teams or journalists?
  2. What is the plan for scaling beyond a single developer?
  3. How does the system handle edge cases where AI fails to extract claims or detect contradictions?
  4. Are there any plans for integrating with existing tools used by investigative teams (e.g., Notion, Airtable)?
  5. What are the long-term goals for monetization and product evolution?
  6. How will the tool ensure accuracy when dealing with ambiguous or incomplete source material?

Back to contents

Investment/Partnership Verdict

The description states:

  • The project is a self-developed prototype submitted to a hackathon.
  • No evidence of traction, revenue, or team size beyond the author.
  • It is not presented as a commercial product or venture.

Not evidenced: There is no basis for assessing investment potential or partnership viability. The tool appears to be in early development and lacks any indication of market readiness or proven demand.

Confidence level: Low — due to lack of external validation, user feedback, or business evidence.

Back to contents

Source

Submitted to the OpenAI 2026 hackathon on Devpost. Project home on DevPost.

The analysis above was generated by a language model from the project's own one-line description. It is not independent research and contains no verified traction, revenue or customer data.