The portfolio, organized by what it solves
What this is: a categorical view of the same 31 BYOK-first products you can browse on the marketplace. Use this lens when you know the problem shape; use the marketplace when you know the product name.
What you do here: jump into the category that matches your problem (BYOK-free releases / voice coaching / business tools / generators / core infrastructure) and clone a product to your own keys.
Why it matters: different operators arrive with different shapes of problem. One lens is the product catalog, another is the problem catalog. Both lead to the same BYOK-first deploy.
BYOK-Free Releases
Public template repos with one-click Vercel Deploy. Bring your own keys; run on your own infrastructure.
Community Question Responder
Vendor community queues sit unanswered for days; meanwhile, third-party answers get moderated out. Both sides need a tool the channel admin can run themselves.
Polls a community Slack or Discord, classifies incoming questions, retrieves from a vendor-specific knowledge base, drafts a high-quality reply in the operator's voice, holds it for one-click approval, posts on approval. Two deployment modes: customer-self-serve (point it at any vendor's public surfaces, drafts for you without posting) or vendor-self-deploy (run it in your own community channel, your team approves drafts). One codebase. Same architecture. BYOK across every key — Anthropic, Supabase, Slack, ElevenLabs, OpenAI, Resend. No CAS-owned fallback.
See product pagePreflight
Drafters discover the same Council-blocking errors set after set; the 5-day round-trip through admin → project owner → AI tools → admin loops the same lessons every quarter.
Drafter drops a PDF; Preflight runs every row of the project's accumulated rejection register against the extracted text and project facts, and returns a one-page triage sheet — confirmed OK / needs eyes / hard fails — before the set leaves the drafter's desk. Deterministic rules for text / logic / cross-reference checks; tight LLM calls only for manual-check rows (Anthropic Sonnet, or OpenRouter alternative). Per-project KB with Supabase + pgvector; Resend for drafter-link emails and inbound planner / engineer correspondence; Google Drive ingestion via OAuth. Every metered API call lands on the operator's account, not ours. Two deployment modes: project-owner-self-host (your own infra, your own register) or drafting-firm-self-host (one deployment across multiple parallel projects).
See product pageGenerators
Platforms that spin up customized white-label versions for a vertical in days, not months.
Rehearsals AI
High-stakes conversations need practice
Our platform generation model for AI voice coaching. Multiple verticals built and ready to spin up for any industry that needs conversation practice.
Visit PlatformRaiseReady Template
Founders need pitch practice, investors need deal flow management
Generator platform for creating white-label versions for founders to get investment-ready and investors to manage incoming pitch outreach. Spin up customized versions in days.
Visit PlatformConnexions
Qualitative research is time-consuming and insights get lost
Generator platform for white-label AI voice interviewers with enhanced AI analysis. Conduct customer research, user feedback, exit interviews, and compliance audits - then let AI extract themes, sentiment, and actionable insights from your surveys and research interviews automatically.
Visit PlatformUniversalLingo
Businesses need scalable multilingual communication
Generator platform for real-time AI translation across industries. One core capability—<1s latency translation in 100+ languages—verticalized into industry-specific products. Spin up white-label translation solutions for tourism, healthcare, government, education, hospitality, and more.
Visit PlatformVoice Coaching
AI voice agents for high-stakes conversation practice and live language assistance.
Kira
Generic AI assistants don't know your context and make you repeat yourself
Not one AI for everyone — a unique thinking partner built around YOUR specific goal. Whether it's career decisions, financial planning, business strategy, or life changes, Kira learns your context through voice conversation and remembers everything. She asks questions before jumping to answers, pushes back when something's unclear, and thinks WITH you instead of just answering.
Visit PlatformBusiness Tools
Vertical platforms solving specific business problems — outreach, deal flow, tokenisation, compliance, and more.
CleanClose
Business shutdown is messy and stressful
Coming SoonF2K Fund Tokenisation
Property investment is illiquid and inaccessible to many investors
Visit PlatformEasy Claude Code
Claude Code terminal sessions are ephemeral — context dies when you close them
Visit PlatformInvestorPilot
Founders struggle to manage investor relationships and deal flow effectively
Visit PlatformHairStylist AI
Clients can't visualize a new cut before committing — and stylists lose time on miscommunication
Visit PlatformF2K Offshore Modular
Australian housing buyers can't safely source offshore modular manufacturers
Visit PlatformFactory2Key Projects
Purchaser-facing project marketing must be cleanly separated from wholesale-investor fund content to manage ASIC/AFSL exposure
Visit PlatformOMQ Outreach
Reaching named procurement decision-makers at scale is slow, manual, and brittle
Visit PlatformCore Infrastructure
Shared trust, security, and observability layers consumed across the portfolio via @caistech/* packages.
Platform Trust
Production AI agents need security, audit trails, and compliance evidence — none of it comes free
The brakes layer for the entire platform portfolio. Six modules: automated security scanning, agent quality evals, structured audit logging, per-tenant token/cost metering, permission governance with human-in-the-loop gates, and rate limiting. Compliance evidence for Australian Privacy Act, OWASP Top 10, and SOC 2 readiness — installed as @platform-trust/middleware in every product.
Visit PlatformPubGuard
AI products ship to enterprises without documented security posture — buyers demand evidence
The security scanning engine inside Platform Trust. Runs on PR merge, deploy events, and weekly cron: dependency audits, secret exposure checks, API endpoint classification (read/write, auth-required), agent tool permission surface analysis, and the OWASP Top 10 checklist for API surfaces. Outputs severity-graded findings, downloadable PDF compliance reports for buyer sign-off, and public compliance badges.
Visit PlatformWant the full product grid?
The marketplace shows every product side-by-side with badges, status, and deploy CTAs.