Updated the waitlist server action to use onConflictDoNothing when inserting records instead of relying on manual Postgres error code catching. This simplifies the logic by gracefully handling duplicate email submissions at the database level and adds necessary documentation for configuring the required DATABASE_URL for local development and deployment.

We've significantly upgraded our SEO and AI-discoverability infrastructure, including a centralized seo.ts utility and new machine-readable llms.txt and pricing.md routes. The rollout introduces new marketing content—such as competitor comparisons and workflow guides—while ensuring our dashboard remains shielded from indexing via optimized noindex and robots.txt configurations.

Conducted a thorough audit and update of the packages/instagram module to ensure API calls and authentication flows remain robust. This maintenance ensures alignment with current API requirements and helps catch potential integration issues early.

We've moved from relyng solely on LLM-inferred values to a more robust approach in lead-value.ts. By analyzing both historical contact memory and defined business offers, the system now intelligently extracts currency mentions and matches them against known price points to provide a more accurate lead valuation.

We updated the system prompts across all Sidekick modes—Lead Analysis, Follow-up, and Auto-reply—to enforce a more natural, professional, and conversational human tone. The prompts now explicitly forbid corporate jargon, fluff, em dashes, and common AI writing patterns like the 'no x, no y, but z' structure. Additionally, we simplified the Lead Analysis output by removing the speculative leadValue field to ensure cleaner JSON responses.

We've replaced static system prompt customizations with a dynamic Supermemory layer, providing Pilot's Sidekick with deeper access to business knowledge and contact history. By leveraging Inngest for background synchronization and backfilling, the Sidekick can now retrieve highly relevant context at inference time, significantly improving personalization. We also deprecated sidekick_setting.systemPrompt in favor of this more scalable approach.

Introduced dark mode-compatible SVG icons for Greptile and Supermemory in the supporters section. Updated the Supporters component to conditionally render these SVGs based on the user's active theme. This improves visual consistency for users browsing in dark mode.

We've removed the informational toast that triggered automatically when the contact sync cooldown was active. This update simplifies the user experience by reducing UI noise during the sync process. Syncing remains restricted during the cooldown period, but users will no longer be interrupted by a persistent notification.

This update transitions our Instagram integration to the Instagram Login API and cleans up significant duplicated logic between server actions and core packages. We also refactored the contact sync system to use an Inngest-first approach with a robust HTTP fallback. These changes ensure more reliable media fetching, better automation matching for specific posts, and tighter billing enforcement.

We've added a new contributors/supporters section to the project's landing page to highlight the organizations helping us build. The implementation uses a reusable component that features interactive tooltips to showcase support details from partners like Greptile and Supermemory. It's a great way to acknowledge the tools and credits that are helping us accelerate development.

We've overhauled the user onboarding experience, now prioritizing immediate value through a new Instagram connection step and an AI-generated DM preview. The flow has been streamlined from 5 steps down to 4 by merging redundant sections, and we've added a robust returnTo cookie mechanism to ensure seamless navigation after OAuth. These changes help users get set up faster and more intuitively.

Cleaned up the Instagram authentication flow by centralizing the cookie name constant and addressed UI inconsistencies in the onboarding form. The onboarding form now correctly handles business type value mapping and input visibility logic. These adjustments fix subtle errors that could have impacted the user's initial setup experience.

Centralized Instagram authentication helpers into a shared utility file, replacing duplicated logic in the auth routes. Additionally, improved the onboarding flow's resilience by adding optional chaining for message username checks and better handling of Instagram preview states during errors. These changes clean up the codebase while making the onboarding experience more robust.

Updated the message selection logic in the onboarding flow to ensure relevant conversations are prioritized. The new implementation properly filters for active business-led threads, while the UI was streamlined by removing extraneous metadata from the message preview cards. This creates a much cleaner and more reliable preview experience for new users.

Updated the Instagram OAuth flow to use a more robust URL generation strategy by prioritizing BETTER_AUTH_URL or NEXT_PUBLIC_APP_URL over derived request origins. This change ensures consistent callback handling across different deployment environments and cleans up temporary URL origin handling. This improves authentication reliability when running behind proxies or custom domain configurations.

Added a new guided onboarding experience for Sidekick to help users configure their offerings, FAQs, and AI tone profile. This multi-step wizard ensures users provide the necessary context for the AI to effectively manage DMs with their audience. The update also includes internal refinements to the Instagram authentication callback logic.