BMAD+ in brief
Described transcript
On screen: Dark blue haze fills the screen. The title “BMAD+ 0.23.0” sits in the centre, and a warm light slowly grows behind it.
Narration: BMAD+ installs roles, workflows and shared context into the AI coding tool you already use.
On screen: A single amber point of light in a dark teal cloud, with faint rays spreading from it. The caption at the bottom reads “Ask first”.
Narration: It starts by asking: a pack declares the questions it needs answered before it acts.
On screen: A terminal window titled “~/app”. A command is typed: bmad-plus intake record, pointing at the Design pack’s intake.yaml and an answers file, with three answers (surface_kind=dashboard; audience=“Accountants in a practice, on a desktop, at month end”; job=UNKNOWN) and a default for design_system. The output reads “Recorded surface_kind, audience, job, design_system”, then “design-read (design): ready — 2 answered, 1 defaulted, 1 unknown, 9 not asked yet.” and a list of the next questions, each tagged with its type (choice, number or text): brand_assets, aesthetic_family, variance, motion, density and stack. The view zooms in on the last line, where “Unknown: job” is outlined: “risk: There is no goal to review against, so hierarchy and the primary action cannot be judged, only decorated.”
Narration: You answer a few at a time, and what you do not know is recorded as unknown, with its risk.
On screen: Amber particles fly outward from a bright centre and form a widening ring against a dark purple background. The caption reads “A swarm, within a budget”.
Narration: Then Nexus can run a swarm of sub-agents, within a budget you set.
On screen: Two terminal panes side by side. On the left, bmad-plus swarm tally buyer-panel-20261018-zho8 is typed. It prints the path of a tally.md file, “8 of 40 units counted; 32 skipped.” and a line that starts with “Disclosure:”: “Simulated units are not people. Read rates as an upper bound and confirm with real ones.” The right pane opens tally.md. Its heading, outlined in orange, reads “Simulated buyers see the pitch (run buyer-panel-20261018-zho8)”, followed by “8 of 40 units gave a valid answer; 32 skipped” and a “Rate: buys” table (the row “all” reads 4 yes of 8, 50.0%). The pane scrolls through short quoted replies grouped by segment and behaviour, then stops on “Numbers: price_floor”: a table with the columns group, count, median, mean, min and max, whose first row, “all”, reads 8, 3.65, 3.7, 3 and 4.5.
Narration: Each one reads a brief and writes an answer, and only valid answers are counted. The tally on screen is demo data, and it says plainly that simulated units are not people.
On screen: A tangled knot of thin amber lines straightens, line by line, into a square grid with a bright point at its centre. The caption reads “Independent checks before acceptance”.
Narration: For planned work, Nexus requires independent checks before it accepts a result.
On screen: Tall, dark, glassy slabs stand in a row, lit from the side by a warm glow. The caption reads “10 packs · 7 adapters”.
Narration: Core's four roles, nine optional packs, and adapters for seven AI coding tools.
On screen: A terminal. The command npx bmad-plus install --lang en --packs core,design --tools claude-code is typed. The installer’s lines appear: “BMAD+ Installer v0.23.0”, “Installing to: /home/dev/app”, “Selected packs: Core, Design”, “Detected IDEs: Claude Code”, a first-name prompt (Developer), the communication language (English), a prompt for a test environment URL, and the execution mode, with Manual, Autopilot and Hybrid offered and “Manual — You call agents yourself” kept. The view zooms in on two green checks: “5 agents, 9 skills, 4 files copied”, with “files copied” outlined in orange, and “1 IDE(s) configured”.
Narration: Install it with one npx command from your project folder.
On screen: A dim blue-black sky with a faint band of light at the horizon. The address “bmad-plus.rochetta.fr” fades in, in orange, and fades out.