In most product teams, context and key decisions live everywhere across Granola call notes, Linear tickets, Figma files, Slack conversations, Notion docs and Github. And with most people adding AI to individual tasks, the result is each tool only sees a fragment of the picture. People get more done, but they don't ship any faster: more code is written, but it's stuck in review; more specs get created but they start from the wrong idea; more insights are created but not acted on.
Enter Warp, a tool we build for every product we work on. It keeps the whole team working from the same product context and provides the foundations AI needs to work well and help teams deliver faster.
What is Warp? Warp’s ‘memory’ is a shared markdown document repository summarising and linking all of the product’s accumulated context, decisions and knowledge. Its ‘connections’ are all the MCP-enabled tools that the team uses - Granola, Jira/Linear, GitHub, Figma, Slack, Notion and Google Drive are just a few common examples. An LLM-driven conversational interface makes the brain interactive - readable, curatable and extensible by all members of the team. All context is wired up via the connected tools - think meeting transcripts, delivery tickets, code repositories, design files, conversations, documents and more.
Beyond just hooking up an AI assistant to various MCP connectors - which is valuable in itself but now common practice - Warp steps up a gear by:
🪮 Combing the context: everything we know about the project - including the product vision, technical architecture, domain knowledge, people, goals, and decisions - is distilled into high-value, concise, well-structured and version controlled documentation. Easy for humans and LLMs alike to navigate quickly.
🪡 Threading through to the sources: The faster-moving granular detail stays in the tools that own it - acceptance criteria in Linear tickets, code implementation in the GitHub repos, agreed decisions in the Granola transcripts. Warp's documents hold the addresses, not the content - file IDs, board and widget links, folder IDs - so a session can pull the live detail on demand and never work from a stale copy.
📝 Following a pattern: the rules that teams follow are written down and version controlled next to the docs, so they govern everyone's sessions rather than living in one person's prompting habits. This includes the small things: defining which source wins when two disagree; agreeing what never gets written into committed repo docs; rules that prevent anyone from creating or changing anything in a central tool without approval first. It also supports the bigger repeatable jobs, packaged as named skills anyone on the team can invoke: turn this week's goals into user stories, clear the triage queue, answer a product-analytics question within the agreed anonymisation rules.
🗑️ Knowing what to throw away: connectors can only show what exists. They can't tell you which document has been superseded, which diagram isn't reliable yet, which folder is off limits, or which recurring transcription error to correct for. Warp records the exclusions and the precedence rules alongside the content, which is often the difference between a plausible answer and a correct one.
🧶 Weaving in what's new: every session begins by checking the connected sources for anything that has landed since the last one - new meeting notes, channel activity, changed delivery milestones - and proposing the specific documentation updates that follow, which are only written back to the shared repo once a team member approves them.
Driving at Warp Speed Warp doesn’t just provide immediate and reliable answers about the product. Because it already holds the context, it helps us drive actions, produce artifacts and push delivery forward at a pace, and a standard, we couldn't achieve without it - and always with a human in the loop.
Tasks it has played an instrumental role in from recent Loomery client engagements include:Planning the work: Running story mapping to establish the roadmap and milestones, then turning each week's goals into granular Linear stories with acceptance criteria written in the style the team signed off on - including the deliberate edge cases that make an AC testable.Shaping the technical direction: Helping define the target architecture, including an entirely new CI/CD and infrastructure setup, grounded in the as-built system and captured domain decisions rather than in best-practice boilerplate.Packaging the recurring jobs as skills: Setting up a new cycle, clearing the triage queue, designing and improving analytics dashboards - repeatable work that draws on Warp's content and runs the same way whoever invokes it.Keeping the delivery record honest: Updating the status of delivery initiatives and tickets in Linear to reflect real progress based on reading the current state of the actual codebase and tracking scope changes agreed in meetings, so the plan of record matches what was actually said and shipped.Writing code in a complex domain: Our client's domain experts share their knowledge into the repo docs and meeting transcripts, and implementation plans then draw on them directly to guide integration with proprietary systems. Understanding a complex domain well enough to write code unaided can take weeks or months. Warp doesn’t replace the expertise; it means the expertise only has to be explained once.
How does it work? Warp is stored as a Git repo with full version history. A team member clones it once, then an AI assistant running inside the cloned folder has immediate access to the shared project context automatically. A new Warp brain is fired up using a script that generates the initial scaffolding for the brain. This spins up a new repository bundling two generic skills out of the box: /brain and /onboard. The latter skill is used once to get the agent to help seed the brain with relevant sources, context and tools. An ingestion ritual means every new session checks all sources for anything new and proposes updates, reviewed and approved by a person, not auto-committed. Warp’s knowledge compounds over time, staying current as a side effect of normal working, without requiring a separate maintenance task. The agent instructions include: CLAUDE.md - an operating guide for Claude - to orient it on the engagement and route it to the live sources; AGENTS.md - which is always loaded and covers quick facts, repo layout and hard rules; and a single shared playbook every tool harness’s adapter points at rather than duplicating instructions per tool. A dashboard app provides a top-level view of the latest developments, where the delivery sits on the timeline and who's working on what. A per-product whitelist file scopes exactly which folders, files and tools an agent can pull from. This matters because the connected sources are often shared across teams - strong access control ensues that the wrong data never gets pulled in. We've built a mock version you can explore , based on a fictional project.
Helping us move fast Consider a real instance where one of our engineers asked Warp to explain the context behind a past incident where one user flooded the system with thousands of files at once. An engineer needed to know what happened and to assess how recent code changes would handle the same situation differently.
That question needs incident history, the original decisions and the current implementation all at once. The information existed but was spread across the original discussion captured in a Granola transcript, resulting in work in Linear, designs in Figma and implementation in GitHub. You won't get that from reading code alone, and answering it the old way meant finding the relevant piece of information in each place and connecting them by hand. Warp pulls all of those threads together at once, closing the information gap instantly.
Our teams are increasingly using Warp to move faster, and when an engagement ends, we transfer the brain repository to the client, so everything the team has learned stays with them.
One tool in our move to become AI native We don’t think AI should just mean doing more, faster, be that writing more code, drafting user stories faster or turning around design iterations quicker. Across the whole team, its greatest strength is in giving us more time to think by offloading time-consuming tasks so bandwidth goes towards higher-level thinking. Warp is one part of how we're building that muscle.
The name, of course, comes from weaving. A warp is the set of threads that give a loom its structure, the backbone the rest of the fabric gets woven into. Fitting, for the thing holding a project's structure together while everything else gets built on top of it.
If you're interested in how Warp could accelerate delivery for you, get in touch .
