Volume Discounts · Product requirements
This file is empty. Write it yourself, or let Releavo draft it.
Idea to release, in one loop
Everything in between, requirements, design, code, QA and delivery, runs in one loop your whole team and its AI share. Releavo drafts every doc and story, follows every push and carries each change down the chain. The right person signs it off.
Mai talks the idea through. Releavo turns it into a PRD draft.
Volume Discounts · Product requirements
This file is empty. Write it yourself, or let Releavo draft it.
1 The product owner approves the brief, PRD and stories. The tech lead approves architecture and the readiness check. QA passes stories. Releavo approves nothing, and the database holds it to that.
The AI does the legwork. People make the calls.
Releavo doesn't sit at one step of your process. It drafts the plan, breaks it into stories, watches the code land, checks it against what was agreed, and forecasts the release. At every step the same rule holds: the AI does the legwork, a named person decides.
Requirements, designs, code activity, tests, decisions and approvals usually live in six tools and three people's heads. Releavo pulls them into one loop and keeps it current, so your team, your client and every AI agent work from the same context. Not a copy of last month's.
Built for the long runA product lives for years and people move on. The record stays: why tiers cap at 500 units, who agreed, and which merge request built it.
Built for teams that hand overAgencies and outsourced teams keep the client in the loop with a read-only view, and every sign-off on record. When the work changes hands, the context goes with it.
A project runs on a chain of documents, each built from the one before. When CR-14 changes the pricing model, Releavo follows your change-request playbook down the chain, drafts every update, and tells each owner what's waiting for them. Nobody goes hunting for what went stale.
Built fromEvery version records what it was made from. PRD v6 says CR-14 + PRD v5, so a reviewer sees why, not just what.
Told, not broadcastOwners hear about it in their own Home and chat: Huy about architecture, Sara about the test cases that moved. Nobody else gets pinged.
Each AI edit lands as the next version of the doc, with a diff, a note on why, and the name of the person who approves it. Approved versions are never overwritten. Roll back to v5 any time.
Wholesale buyers get a flat 10% off orders over 50 units up to three tiers, set per customer group: 5% from 50 units, 8% from 200, 12% from 500.
Tiers apply to the whole order, after any customer-group price. Discounts do not stack with promo codes. A promo code replaces the tier discount if it's larger; they never stack.
The cart shows the discount applied, the tier reached, and how many more units unlock the next one.
Enforced in the database, not in a prompt. If your roles can't approve a doc, no button, script or clever request will.
Connect the AI tools your team already uses (Claude Code, Cursor, VS Code, ChatGPT, anything that speaks MCP) and they work straight from the story: the criteria, the epic, the architecture, moving the story as they go. Or switch coding on and Releavo builds it itself, as a draft pull request under your name. Either way, every push and merge request reaches Releavo, which finds the story, moves it, and checks the code against what was agreed.
releavo/VD-35-cart-tiersreleavo/ branch, under the name of whoever pressed it. It never touches another branch or merges.feat/VD-35-tier-ids → maintotal.ts still adds the promo on top of the tier.031_tiers.sql isn't in architecture v3. Told Huy.Designs stay in Figma and show up live where they're needed. No designer yet? Releavo sketches a clickable prototype. Drop in the spreadsheet the idea came from, or talk instead of typing. All of it lands in the same loop.
Connect Jira and Releavo looks around, recommends what to bring in, and asks before moving anything. From then on it syncs in the background: approved here goes out there, changed there comes back. Any MCP server works too.
Sign-ins stay on Releavo's servers, encrypted. The agent calls your tools through Releavo and never sees a token, and client stakeholders can only read.
Epics and stories, sprints you plan and close, velocity from what your team actually finished, and a release forecast that moves when scope does. Brought in from Jira or built here.
Roles belong to people, per project. Mai is PO and BA on Volume Discounts and a client on Brewline. Releavo works with exactly the roles of whoever's asking, never more. Pick someone:
Split VD-35 into a cart story and a checkout story.
Drafted VD-35 v3 and a new VD-44 for you to approve. Told Huy both need estimating.
There's no change-request button. There's playbooks/change-request.md: plain markdown your team writes and edits. Releavo reads it the way a new hire reads the wiki, then follows it. Change the file and you've changed the process.
change-request.mdWrite it up, trace the impact, give the PO optionsreadiness-check.mdIs the architecture ready for stories to start?stories.mdStories, acceptance criteria, traceability to the PRDtest-plan.mdTest cases from acceptance criteria, bug reportsdoc-chain.mdAfter an approval, bring every later doc into stepstatus-update.mdThe weekly note a client can actually readrelease-notes.mdFrom done stories to release notesuploads.mdRead workshop notes, file them, pull out decisionsEach job gets its own sandbox with the project's files in it. Releavo reads and edits them, runs Python on a Word doc or a spreadsheet, searches the web, and hands focused work to specialists. Then everything it changed comes back as drafts.
Word, Excel, PowerPoint and PDF too, and clickable HTML prototypes. It edits a copy of your files, and the copy is a draft.
Everyone gets a Home built from their roles: what's waiting on their approval, which stories are blocked on them, what they've been asked. Not a feed of everything that happened while they were away.