vs Archistar
Archistar generates the massing. Feasibly verifies the rules the massing must obey.
Archistar is a generative-design platform: AI massing, design-led feasibility, built-form visualisation. That is design tooling, and it is genuinely good at it. But every design starts from a planning envelope, and that input has to be right. Feasibly is the verified constraint feed: the schedule-specific envelope, the permit triggers, and the yield verdict, cited to the clause, confidence-tagged, underwritten, in seconds.
Head to head
Two different jobs, and the one that comes first
Archistar sells the design layer. Feasibly sells the input layer: the interpreted planning constraints a design has to live inside. The comparison below is about that first job only.
| Archistar | With Feasibly | |
|---|---|---|
| Core job | Generative design: produce massing options and design-led feasibility studies for a site. | Interpret the planning scheme: the envelope, the permit triggers, and the buildability verdict for an address, before anyone designs. |
| Schedule-specific envelope | Surfaces standardised planning controls in its property layer. We found no published evidence of council-schedule-level depth (the 620+ Victorian schedule variations). | Per-council schedule interpretation plus the ResCode Clause 55 envelope engine, current to VC267: height, setbacks, coverage, private open space. |
| Source citation & confidence | States its data comes from government sources. We found no inline clause citation or per-value confidence score. | Every value carries its source clause URL and a verified / machine-extract / default confidence tag. |
| Permit triggers | Compliance-checks a finished design against controls, a different, later job. We found no forward-looking “does X trip a permit here” output. | Answers it before you design: 2,306 overlay-schedule permit rules across all 79 VIC councils, plus NSW consent triggers, with the exemptions to check. |
| Buildability / yield | Yes. Design-led: generate schemes and study feasibility. Strong where you already intend to design. | Screening-grade: subdivision estimate plus the likely assessment pathway (VicSmart / standard / referral) as a go/no-go before design spend. |
| Accuracy posture | We found no published accuracy figure or guarantee. | Published benchmark (100% zone accuracy on the cross-state set) and a guarantee: wrong zone → free lookup + 10 credits + a public dataset correction. |
| Time to the answer | Open the platform, set up the site, generate options. | One lookup. Seconds, in the web app, the API, or inside Claude / ChatGPT / Cursor via MCP. |
| Economics for per-address checks | $95–$595/month per seat, priced as a design platform you live in. | Credit-metered: Free 20/mo → $29 for 200 → $99 for 2,000. Priced as an input you consume dozens of times a week. |
Decision
When to reach for each
Reach for Archistar when you need to…
- Generate massing options for a site you’ve committed to
- Run design-led feasibility studies with built form
- Visualise and iterate schemes in 3D
- Compliance-check a design you’ve already produced
Reach for Feasibly when you need to…
- Get the cited, schedule-specific envelope before pen meets paper
- Know what trips a permit here, before you design, not after
- Screen many addresses fast for a go/no-go verdict
- Feed verified constraints into your design workflow, including Archistar’s
Closing
Our honest position
- Archistar is a real design platform, and generative massing is craft tooling we don’t compete with and never will.
- The job before design (the verified envelope, the permit triggers, the yield verdict) needs different machinery: clause citations, confidence tags, schedule-level codification, an accuracy guarantee. We publish all four; we found no published equivalent in the design platforms.
- We publish our accuracy and underwrite it. That’s not a claim about anyone else’s data. It’s a standard you can hold us to.
- Use both: Feasibly is the input layer, Archistar is the design layer. MCP + REST means the feed drops into whatever you design in.