PlanMerge

What the company appears to be: PlanMerge is a self-reported tool designed to merge fragmented AI-generated planning drafts into one structured, decision-traceable team plan. The author states it normalizes inputs from multiple AI tools, merges them into "Decision Blocks", and preserves source excerpts, rationale, and human overrides.

What changed: During the OpenAI Build Week hackathon, the project was extended with Codex and GPT-5.6 to introduce a new "Decision Room" workflow that generates clarifying questions for unresolved conflicts and re-merges sections accordingly.

The single most important open question: Is there any evidence of real-world usage or adoption beyond the author’s own development and testing?

The description states that PlanMerge:

• Turns multiple AI-generated planning drafts into one structured, reviewable team plan;

• Normalizes drafts from multiple AI tools into canonical ideas;

• Merges them into section-level "Decision Blocks";

• Shows selected options, rationale, alternatives, conflicts, and exact source excerpts;

• Scores section coverage, source coverage, and decision traceability through a "Quality Gate";

• Lets teammates vote, leave anonymous opinions, override AI choices, and preserve the change in a "Decision Log";

• Exports the final plan as Markdown or portable JSON workspace.

It is built using Next.js 16, React 19, TypeScript, deployed on Vercel. The system uses a two-stage pipeline: one for normalization and another for generating structured Decision Blocks. Server-side validation cross-checks enums, IDs, and citations; it retries malformed output with repair prompts and falls back to an offline deterministic harness when no API key is available.

Shared workspaces use Neon PostgreSQL and Prisma. It supports local-first functionality for fast testing, with shared links, anonymous voting, link expiry, revocation, rate limiting, quality regression cases, and Playwright end-to-end coverage.

Codex was used to help define reliability invariants, implement API/UI flows, create quality/e2e harnesses, and audit failure paths across the merge pipeline.

Not evidenced: The actual product interface or user experience beyond what is described. No screenshots, demos, or live links are provided.

The author states that PlanMerge was built to make the decision process visible when teams use AI tools like ChatGPT, Claude, and Gemini to draft plans from different roles. The result of such fragmented inputs is not a better plan but a pile of plausible documents — which then must be manually merged by hand.

This leads to loss of visibility into:

• Which ideas were selected;
• Which alternatives were rejected;
• Where opinions conflicted;
• Which source supported each decision.

PlanMerge aims to solve this by making the merge process transparent and traceable.

The author also claims that disagreement is valuable data, and a strong planning tool should not silently average conflicting proposals but help teams understand and resolve them.

During OpenAI Build Week, the project was extended with Codex and GPT-5.6 to add a "Decision Room" workflow that generates clarifying questions for unresolved conflicts and re-merges sections accordingly.

Not evidenced: The positioning relative to existing tools or platforms (e.g., Notion, Confluence, Airtable). No competitor comparison or market differentiation is provided.

The description states that PlanMerge targets teams who use AI tools like ChatGPT, Claude, and Gemini to draft product plans from different roles. These teams often end up with multiple drafts that must be manually merged, leading to loss of decision traceability.

It also implies a need for structured workflows where:

• Team members can inspect evidence;
• Vote on choices;
• Override AI decisions;
• Export reasoning alongside the polished document.

Not evidenced: Specific customer personas or use cases beyond general "teams using AI tools". No segmentation or targeting beyond "teams" is described.

The description does not contain any information about pricing, monetization strategy, or business model.

Not evidenced: No evidence of revenue streams, pricing tiers, subscription models, or customer acquisition methods.

PlanMerge is built using:

• Next.js 16 and React 19 with TypeScript;
• Deployed on Vercel;
• Uses Neon PostgreSQL and Prisma for shared workspaces;
• Playwright for end-to-end testing;
• Codex for turning product principles into working code.

It implements a two-stage pipeline:

1. Normalization of drafts from multiple AI tools.
2. Generation of structured Decision Blocks.

Features include:

• Server-side validation of enums, IDs, and citations;
• Retry mechanism for malformed output using repair prompts;
• Fallback to offline deterministic harness when no API key is available;
• Local-first design for fast testing;
• Anonymous voting and opinions;
• Link expiry and revocation;
• Rate limiting;
• Quality regression cases.

Not evidenced: No information on scalability, performance metrics, or production deployment details beyond the development stack.

The description states that:

• A one-click verified sample loads 13 role-based drafts and produces a 12-section plan with complete source coverage and an intentional MVP-scope conflict;
• Judges can test the core value immediately without providing credentials;
• The project existed before OpenAI Build Week but was meaningfully extended during the submission period.

Not evidenced: No evidence of real-world usage, customer feedback, or adoption beyond the author’s own testing. No metrics on user engagement, retention, or product usage are provided.

The description does not provide any information about existing competitive products or platforms in the space of AI planning tools or decision traceability.

Not evidenced: No mention of competitors such as Notion, Confluence, Airtable, or other collaboration or planning tools. No market positioning or differentiation is described.

• No real-world usage: The project appears to be a prototype built for a hackathon with no evidence of adoption or traction.
• Unverified claims: All features and functionality are self-reported without independent validation.
• Limited scope: The tool seems tailored for AI-generated planning drafts, which may limit its applicability beyond this niche.
• Dependency on AI APIs: The system relies heavily on external AI services (e.g., Codex, GPT-5.6), which could pose risks if those services change or become unavailable.
• Unclear monetization path: No business model or revenue strategy is described.

1. What specific problems are teams currently facing when merging AI-generated planning drafts?
2. How does PlanMerge handle disagreements between AI models or team members?
3. Is there any feedback from early users or internal testing beyond the author’s own experience?
4. What is the long-term vision for monetization and product development?
5. Are there plans to support more than just Korean service plans, or expand into PRDs and business plans?
6. How does PlanMerge ensure data privacy and security in shared workspaces?
7. What are the technical limitations of the current architecture that might affect scalability?