Harth

Overview

Harth brings architecture project work, model context, visualization and team collaboration into a browser workspace, with AI agents that work on building data. It is best suited to architecture practices that want these activities connected around shared projects, rather than a standalone rendering app.

The breadth is its strongest argument: every plan includes the full workspace, while the central trade-off is how much AI usage its monthly credits cover.

Key features

Harth imports Revit, Archicad, SketchUp and IFC models while preserving elements and coordinates. That gives practices a way to bring existing project models into a shared environment without recreating them. Its roughly 40 discipline-specific agent skills span areas such as structure, envelope, interiors and analysis; this is a wide remit, but credit consumption will determine how much a team can use the AI tools.

Site tools combine 3D satellite terrain, surrounding-building imports, boundary tracing, geolocation and sun studies. They make Harth more relevant to site and early-design work than a renderer alone, though teams seeking only image generation may not need the wider workspace.

For visualization, users can frame a view in the model, mask areas, apply lighting presets and upscale images for print. Collaboration includes simultaneous model work, live cursors, presence, threaded comments pinned over 3D and per-project voice, making the workspace suited to shared review rather than isolated production.

Harth also says 360-camera footage can become an explorable, measurable 3D scan, processed in one to two days. Exports include PDF, image and 3D files, 360-degree views and recorded walkthroughs, supporting both conventional deliverables and immersive presentations.

One important consideration is AI data handling: Harth NZ Limited says prompts and generated outputs may be stored for platform improvement and quality assurance, and anonymized or aggregated prompt data may be used to develop and train its AI systems. Its privacy policy addresses New Zealand and Australian privacy laws and describes rights for EU and UK individuals under GDPR.

Pricing

The Free plan costs nothing, billed $0 per month, and includes 700 credits / month, 20+ renders and community support. It is a sensible way for an individual or small team to assess the workspace, but the modest credit allowance limits sustained AI use.

Plus costs $59 a month, billed $59 per month, for 5,000 credits / month, 150+ renders, priority rendering and email support. It is the entry paid tier for regular use. Pro costs $149 a month, billed $149 per month, and raises the allowance to 15,000 credits / month and 500+ renders, with priority rendering and priority support; it better fits heavier AI workflows or practices that value faster escalation.

All plans include the same workspace, so the higher subscription tiers buy more AI capacity and support, not a different feature set. Credit packs offer a one-time alternative: Small is $99, paid once for 7,500 credits and 250+ renders; Medium is $199, paid once for 16,000 credits and 500+ renders; Large is $399, paid once for 36,000 credits and 1000+ renders. Packs never expire and suit occasional bursts better than upgrading monthly, though they do not replace a recurring allowance. Monthly credits can also be topped up with packs. A free trial is available.

Platforms

Harth runs on the web, so teams can use the browser-based workspace rather than a desktop-only application.

Who it's for

Harth says it is built for architecture practices of roughly 15 to 30 people in New Zealand, Australia and the United States. Its combination of model imports, site tools, shared review and multiple export formats makes the strongest case for practices coordinating design and presentation work across a team. It is less compelling for someone who needs only a narrowly focused rendering tool.

Pros and cons

• Pro: Existing Revit, Archicad, SketchUp and IFC models retain elements and coordinates on import, reducing the case for rebuilding project data.
• Pro: One shared workspace covers site context, model-based visualization, collaboration and exports, while every plan retains the full workspace.
• Pro: Credit packs never expire, giving teams a way to handle occasional high-demand periods without changing their monthly tier.
• Con: AI usage is governed by monthly credits, so the workspace's broad tools do not mean unlimited AI work.
• Con: Prompts and generated outputs may be stored and anonymized or aggregated prompt data may be used to develop and train AI systems; teams with strict data requirements should review that policy carefully.

Alternatives

For a dedicated AI architectural rendering category, browse AI Architectural Rendering Software. Choose Luw.ai Studio if iOS or macOS access matters alongside web; its free tier includes 2 credits every day but limits output to 512×512 with a watermark. Gendo is another web-based freemium option, with a free plan that allows unlimited generations but restricts users to one collaborator and watermarked output.

Treform is worth considering if macOS or Windows support is important in addition to web access. mnml.ai offers an API as well as a web platform, making it a better fit when API access is wanted. Renderica has a free allowance of 3 renders per month at 1024px, which may suit light rendering needs.

Archome AI is a paid option with a free trial and an API, extension, macOS, web and Windows platforms. Rendero is a web-based freemium alternative whose free allowance is a one-time 50 credits rather than a monthly refill. Massing Labs is another paid, web-based option.

Verdict

Harth is the stronger fit for a mid-sized architecture practice that wants model imports, site context, team review and presentation tools in one browser workspace. Its full-workspace plans keep the choice focused on AI credit needs, while never-expiring packs suit occasional spikes. Look elsewhere if you need only rendering, or if the stated prompt and output data handling does not fit your project requirements.