OzigiOzigi

Release notes

Changelog

What's new in Ozigi — every update, shipped and documented.

  1. v1.0

    Image Upgrade, Architecture & June Content Push

    Image generation moves to the stable Gemini 3.1 Flash Image model, the architecture page is replaced with current system documentation, the gating system is simplified, and eight long-form GTM articles ship across the blog.

    • AIImage generation upgraded from `gemini-3.1-flash-image-preview` to the stable GA model `gemini-3.1-flash-image`. No change to the server-side R2 upload pipeline.
    • FixCredit bundle return URL (`?checkout=credits`) now fires the Google Ads conversion event, shows a 'Credits added' toast, and cleans the URL param — it previously landed silently with no feedback.
    • ImprovementArchitecture page rewritten with five current decision records: GTM pipeline, JSON schema enforcement, image generation pipeline, and the Banned Lexicon (including the LinkedIn 360Brew compliance connection). Removed the dated LLM cost-comparison tab.
    • ImprovementTrial gating replaced — TrialBanner and TrialGateModal removed. LimitModal is now the single surface for plan-limit nudges, without the time-pressure framing.
    • FeatureGTM blog series: eight long-form articles on go-to-market for developer products published across June 1–12. Covers ICP definition, GitHub/Dev.to lead sourcing, email vs LinkedIn, cold email copywriting, daily send limits, and domain warmup.
    • InfraREADME fully rewritten to reflect the current two-engine product, updated pricing, and accurate tech stack (R2, Dodo Payments, Composio CRM, Gemini 3.1 Flash Image).
  2. v0.9

    Pricing Overhaul & Dodo Payments

    The entire pricing model is replaced. The old Team/Organization tiers are retired in favour of a five-tier system built around the two engines. Payments move to Dodo Payments with a full webhook pipeline into Supabase.

    • FeatureNew pricing tiers: Free ($0 · 50 GTM credits + 3 content pieces), Starter ($19 · content engine only), Growth ($29 · GTM engine · 1,000 credits/mo), Pro ($49 · both engines, no limits), Enterprise (custom).
    • FeatureCredit bundles for Starter users who want outbound without upgrading: 200 credits for $5, 500 for $10, 1,500 for $25. Credits stack on the monthly plan and never expire.
    • InfraPayments moved to Dodo Payments. `/api/create-checkout` handles subscription plans; `/api/create-bundle-checkout` handles one-time credit bundles. Both call the Dodo live API and return a `checkout_url`.
    • InfraDodo webhook at `/api/dodo-webhook` processes subscription lifecycle events and writes plan state directly to Supabase.
    • ImprovementMulti-inbox rotation added as a Pro-only feature. Campaigns on Pro can spread daily sends across multiple sending accounts to protect domain reputation at higher volumes.
    • ImprovementContent entitlements restructured: newsletter sending moved to Starter+, image generation (2/campaign) added to Starter, scheduling moved to Starter+, subscriber list management and campaign analytics gated to Pro+.
  3. v0.8

    GTM Engine — Lead Sourcing, Scoring & Outreach

    The GTM engine ships as a full second product line inside the same dashboard. Source leads from GitHub, Dev.to, and LinkedIn, score them against an ICP with Gemini, run email and LinkedIn sequences from your own accounts, and sync to your CRM — all in one voice.

    • FeatureLead sourcing from GitHub: bio-keyword + language + location query against the GitHub user search API. Commit-history email recovery when a profile hides its email address.
    • FeatureLead sourcing from Dev.to: pulls authors by tag, matched against ICP topic keywords.
    • FeatureICP scoring: every sourced lead is scored by Gemini (0.0–1.0) against the ICP defined on the campaign. Leads below threshold are dropped before entering any sequence.
    • FeatureEmail + LinkedIn sequences: multi-step outreach from your own accounts with configurable step delays and per-channel daily limits. Reply detection pauses the sequence automatically on any inbound response.
    • FeatureCRM sync: first-contact write to HubSpot, Zoho, or Salesforce via Composio. Available on Growth and Pro.
    • AIGTM outreach copy uses the same Banned Lexicon and persona system as the content engine — the cold email and the blog post read as the same person because they are written by the same engine.
    • InfraGTM cron workers (`/api/gtm/cron/scrape` and `/api/gtm/cron/send`) handle sourcing and sequence delivery on schedules managed by QStash.
  4. v0.7

    Generation Reliability & Media Support

    A full reliability pass on the generation engine — output quality guardrails, faster file handling, three production bug fixes, and Google Ads tracking.

    • AILexicon quality guard — generated campaigns are now validated against a banned-phrase list (AI clichés, filler structures, weak closers). Flagged drafts are automatically repaired with a second targeted call before being returned to the user.
    • ImprovementFaster file & image generation — images, audio, and PDFs are now passed to Vertex AI as direct HTTPS URIs instead of being fetched and base64-encoded, cutting per-file overhead and keeping requests well within the 60s runtime limit.
    • ImprovementFile upload warning — an amber notice now appears whenever attachments are added, advising users that larger files increase generation time.
    • FixComposio OAuth redirect — connecting integrations no longer redirects to localhost in production. The callback URL is now derived from the live request origin.
    • FixPersona manager infinite spinner — the loading state now always resolves, even when the Supabase fetch errors or returns no data.
    • FixGeneration stuck at loading — restored the synchronous generation flow that was silently failing in production due to a misconfigured async worker pipeline.
    • InfraGoogle Ads conversion tracking (AW-18111303438) wired up across the main app and blog via a shared `gtag` module.
    View PR #117
  5. v0.6

    LinkedIn & Gemini 3 Upgrade

    Direct LinkedIn publishing arrives and the generation stack moves to Gemini 3 Flash, with a landing-page demo and a handful of stability fixes.

    • FeatureLinkedIn engagement — direct LinkedIn publishing and engagement support alongside existing X/Twitter integration.
    • AIGemini 3 Flash upgrade — generation routes moved from Gemini 2.5 Flash to `gemini-3-flash-preview`. Image generation remains on 2.5 Flash for stability.
    • Infra`getVertexAIClient` moved to global scope to support Gemini 3's initialization requirements.
    • FixLong-form content history now persists correctly.
    • FeatureInteractive demo added to the landing page.
    View PR #106View PR #107View PR #108
  6. v0.5

    Long-Form Content & Marketplace Personas

    Long-form writing lands with full persistence, the persona marketplace is organised for discovery, and the first technical blog post goes live.

    • FeatureLong-form content generation — blog posts and technical docs with database persistence and word-count tracking.
    • ImprovementMarketplace personas — eight pre-built writing personas ordered with `order_index` for consistent display; JSON parsing issues resolved.
    • FeatureFirst technical blog post published to `/blog` — covers Ozigi's retrieval-augmented generation architecture.
    • FeatureSubscription management — upgrade and cancel plans from the dashboard; org-level rate limits set to 5 requests / 24h.
    • ImprovementBlog SEO — meta tags, structured data, and sitemap improvements across all blog pages.
    • FixBuild path errors resolved using `process.cwd()` for reliable cross-environment deployment.
  7. v0.4

    Blog, Analytics & Landing Page

    The blog infrastructure ships, the landing page gains interactive comparisons and a testimonial carousel, and site analytics are in place.

    • FeatureBlog section launched at `/blog` — five categories (Engineering, Marketing, Content, Tools Roundup, Ozigi Focus) with author, read time, and category metadata.
    • FeatureDrag-to-reveal before/after slider on the landing page's "difference" section.
    • FeatureTestimonial carousel — endless-scrolling with six entries, seamless animations, and edge fade gradients.
    • InfraVercel Web Analytics integrated for site analytics.
    • ImprovementAll outbound emails now use `hello@ozigi.app` as the sender address.
    • ImprovementDashboard loading states — improved skeleton loaders and confirmation dialogs.
  8. v0.3

    Dashboard Redesign & Payments

    A full dashboard redesign, streamlined upgrade flows, and a more polished AI Copilot — plus end-to-end tests for the core journeys.

    • ImprovementDashboard redesign — cleaner information hierarchy, revamped auth modals, improved error handling.
    • ImprovementAI Copilot — "thinking" state, working copy button, better styling and progress indicators during generation.
    • FixMoved to a stable generation API endpoint — eliminates phantom usage increments and streaming response issues.
    • ImprovementUpgrade flows simplified; demo dashboard redesigned for a clearer conversion path.
    • InfraPlaywright end-to-end tests for core flows.
    • ImprovementSEO & branding meta improvements across documentation and blog pages.
    • InfraInitial Vercel Web Analytics install (superseded by v0.4).
  9. v0.2

    Scheduled Publishing, Project Imports & Email

    Scheduling, GitHub project imports, and the first email automation workflows land — with fixes to plan limits and Vertex AI auth.

    • FeatureScheduled post publishing — pick when content gets published instead of publishing immediately.
    • FeatureGitHub project imports — pull in repositories as context for content generation.
    • FeatureEmail automation — welcome email system and automated email workflows.
    • FixFree plan monthly counter now resets correctly; trial plan detection fixed to assign correct generation limits.
    • FixVertex AI authentication resolved in sandbox environments.
  10. v0.1

    Persona System & Foundation

    The persona system — the heart of Ozigi's voice control — ships with custom persona creation, a settings modal, and the foundational schema.

    • FeaturePersona voice selection — apply writing personas to shape the voice of generated content.
    • FeatureCustom persona creation beyond the system defaults.
    • FeatureSettings modal for managing personas and preferences.
    • InfraDatabase schema — `personas` table renamed to `user_personas` for clarity.
    • ImprovementPersona UX refactor — improved select dropdown and management flow in SettingsModal.
    • InfraApp metadata — page titles, descriptions, and layout foundations.

Have feedback?

Tell us what to build next.

Email the teamReach out on X