Project Lift

Inspiration

On almost every complicated program or project, there is someone quietly carrying knowledge and not so quietly moving the team forward.

You know them. You rely on them. Maybe you are them.

They remember why a decision was made three months ago. They notice when two teams are solving adjacent versions of the same problem. They know which dependency is about to become a blocker, which meeting actually needs to happen, which requirement has quietly changed, who isn’t talking to the right team, and which small problem is about to become very expensive if nobody catches it.

They may be an engineer, analyst, researcher, product lead, operator, or simply the person who has come to understand enough of the whole system to keep it moving.

It is an invaluable human capability, but at its core it is sustained memory, attention, connection, and follow-through.

Project Lift aims to free these individuals to do more of the work only they can do and ask the organization's systems to carry more of that load themselves.

What if every project had a persistent working memory?

What if decisions, dependencies, commitments, open questions, owners, status, changes, and risks did not disappear into meetings, documents, and message history?

What if the system could notice, surface, and resolve a critical thread that has gone quiet and didn't light up elsewhere, that two assumptions conflicted, that an unresolved decision was blocking three downstream tasks, or that something everyone thought was handled…wasn’t?

Project Lift is about giving the organization more memory, more continuity, and more ability to carry its own complexity - so the people inside it have more of themselves available for the work they are truly here to do.

What it does

Project Lift brings scattered updates, documents, decisions and dependencies into a shared picture of the work. It checks new information against its sources, identifies contradictions and blockers, and tracks what needs verification, so executives, teams and individuals can see what’s happening and what needs attention. Every status can be traced back to its evidence, helping people keep the work connected, current and moving without people carrying all that context in their heads.

Built with

HTML, CSS, JavaScript, TypeScript, Node.js, HydraDB, OpenAI Codex

Frontend: HTML, CSS and JavaScript
Backend: TypeScript on Node.js
Knowledge storage and retrieval: HydraDB
On-demand agent reasoning: local Codex CLI
Workflow state: local file persistence

Challenges

The central challenge was turning scattered updates into trustworthy shared context: distinguishing a new claim from an accepted status, preserving conflicting evidence, and making uncertainty visible. -Integrating HydraDB also required separating upload acceptance, searchable content and graph processing. -Live retrieval is verified; the graph’s contribution to application answers remains under evaluation -Multi-agent development exposed its own shared-context problem: implementation, verification and the live preview were operating on different checkpoints. It turns out our agents need Project Lift, too.

Accomplishments

We built a working prototype that combines live HydraDB retrieval with persisted individual updates and an inspectable verification workflow. In a recorded investigation, the agent selected a verification step and escalated when the evidence was insufficient. We added multi-project support, source inspection and evidence freshness indicators, backed by regression tests and independent review. The result demonstrates a foundation for keeping shared work connected, current and moving without hiding uncertainty.

What's next for Project Lift

Add a monitoring agent that watches connected work sources for changed commitments, emerging blockers and conflicting updates, then brings relevant context to the people affected
Extend on-demand triage to ask targeted verification questions, follow unresolved issues and propose evidence-backed status changes for review
Evaluate whether graph-backed relationships improve dependency discovery and cross-team handoffs, especially when sources disagree
Connect production systems and add authentication, access controls and deployment safeguards