Where Mañana Is Going

A platform that grows with you. Data to decisions, email to automation, notes to AI. Here's what's next on the journey.

0
tests passing
0
view system files
0
workflow node types
0
field types
0
AI services
0
microservices

Document Template System

22 March 2026 Planned

What's a company without paperwork? This is a dedicated visual editor for designing document templates. Place fields on uploaded images to create multi-page PDFs that populate themselves with your data. Invoices, contracts, certificates, designed once and filled forever through workflows.

Visual Field Placement

Drag and drop data fields onto page images. Position, resize, and style each field with pixel precision. Preview with live data before saving.

Multi-Page PDF Generation

Compose templates from multiple page images. Each page can pull from different tables and relationships. Export as professionally formatted PDFs.

Workflow Integration

Link template fields to any column in your schema. Trigger document generation from workflow actions so a new row creates a PDF and emails it automatically.

Workflow Completion

25 March 2026 In Progress

Nearly there. Finishing the workflow automation engine, completing the transform pipeline, wiring AI nodes to real language models for document and image parsing, and making sure data flows exactly where it should across all 33 node types.

Transform Pipeline

Map, filter, and aggregate data mid-workflow. Transform nodes reshape data between steps. Rename fields, filter arrays, compute totals, all without code.

AI Document Parsing

The AI interpret node connects to multimodal models. EU-hosted Mistral on free plans, fully local models on paid plans for complete privacy. Extract structured data from documents and images inside any workflow.

Data Flow Confidence

Full test coverage for the data flow resolver. Every node's output schema is statically inferred so you see exactly what variables are available at each step.

Production Hardening

10 April 2026 Planned

Nobody notices infrastructure until it breaks. This phase is about consolidating the platform for reliability. Extracting shared code into a common layer, hardening authentication, improving test coverage, and building the hierarchical API that will power AI-driven automation.

Common Service Layer

Shared authentication, middleware, and utilities extracted into a single package. One bug fix, every service benefits. Consistent auth behaviour everywhere.

API Discovery System

A hierarchical, self-describing API that allows language models to discover platform capabilities. The foundation for MCP-style tool use by AI.

Query & Security Hardening

Optimised database queries, hardened auth endpoints, improved rate limiting. The unglamorous work that makes the difference between a demo and a product.

Every Language, One Platform

After hardening Planned

Your team speaks more than one language, so your tools should too. Type a country code or language name into your settings and AI generates every interface label automatically. If a language exists in any AI model, it works here.

Context-Aware Components

Every UI component carries a description of its purpose, so AI translates by meaning, not word-for-word. Type a language code and the whole interface adapts. No translation files needed.

AI in Your Language

The chat assistant and email secretary adapt per user. Ask questions in French, get French answers. Summarise emails in Japanese, get Japanese summaries. Yoruba, Catalan, Tagalog, it just works.

Per-User, Not Per-Org

Language is a user-level setting, not an organisation-wide toggle. A team across three countries can each use the platform in their own language simultaneously.

Communication Services

Parallel with hardening Planned

People talk everywhere and your platform should keep up. A chat microservice that doubles as an AI conversation interface, unified messaging across WhatsApp, Telegram, and Discord, and workflow-triggerable notifications across all channels.

Chat & AI Interface

A simple inter-user chat system that also serves as the primary AI interaction surface. Ask questions, trigger actions, and get contextual answers.

Unified Messaging

One interface to manage WhatsApp, Telegram, and Discord. Messages appear in a dedicated table, searchable and filterable. No more switching apps.

Workflow-Triggered Comms

Send a Telegram message when a deal closes. Post to Discord when a deployment finishes. Trigger any channel from any workflow with template variables.

Data Visualisation Overhaul

After communications Planned

Data's only useful if you can actually see it. This is a fundamental rethink of how data is viewed, referenced, and composed across the platform. A new reference syntax (RefSys) lets you embed live data anywhere, from notes to messages to documents, and a modular UX makes every view extensible.

Reference System (RefSys)

A universal syntax for referencing data anywhere on the platform. Type @table.row.field in a note and it renders as a live chip showing the current value.

Rich Notes & Documents

Notes graduate from plain text to full rich text, possibly as a dedicated microservice. Markdown rendering with live RefSys embeds so your documentation stays current on its own.

Modular View Extensions

File browsers, email displays, and custom visualisations become composable modules. Plug them into any view layout. The 87-file views system becomes the extensible foundation.

AI Operations Interface

After data overhaul Planned

Most AI tools guess what they can do. Ours knows. The AI becomes a first-class operator of the platform, a unified interface combining the MCP API, RefSys, and chat into an assistant that understands every capability, confirms operations visually, and honestly tells you what it can and can't do yet.

Platform-Aware AI

The AI knows every API endpoint, every table, every workflow. It can explicitly say "I can't do X yet, but I can do Y" instead of hallucinating capabilities.

Confirmation Modals

Before the AI performs any operation, it shows a visual summary of what will happen. A "skip confirmations" option for power users who trust the system.

External Tools

Web searches, public data crawlers, and third-party integrations available as AI tools. The assistant reaches beyond the platform when your question requires it.

The Long Horizon

The full picture. Mañana as a complete operational platform. These are the ambitious goals that drive every decision we make today.

Hosted Websites

External auth and fully custom HTML/CSS/JS layouts. Your views system becomes a website builder with user signups and public access.

Document Editing

LibreOffice extension compatibility for browser-based document editing. An open-source alternative to Google Docs, with import/export for PDFs.

Database Connectors

Direct PostgreSQL tunnels for external systems. Python libraries that let you use Mañana as your actual database layer in custom integrations.

Voice Input

Whisper-powered voice-to-text for creating notes, triggering workflows, and dictating data. Eventually a mobile keyboard app for on-the-go workflows.

Zero-Knowledge Fields

Encrypted fields where the platform genuinely cannot read the data. Frontend-only encryption with user-held keys. True privacy by design.

Mobile Keyboard

A simple iOS/Android keyboard app connected to Mañana. Dictate notes, look up data, and trigger workflows without opening the full app.

Tomorrow starts here.

Mañana is being built in the open. Follow our progress, or get involved.