My role
Conjoin is the deeply technical side of my work at Delalify. I architect and build its core infrastructure, networking layer, and distributed systems. I designed the service routing, multi-region deployment topology, and cross-service coordination that hold eight independent services together, and I work on them directly. As Delalify's founder and principal engineer, this is the part of the work I stay closest to, building the systems the platform runs on while I advise the rest of the business part-time.
Conjoin replaces the pile of vendors you normally connect by hand when building SaaS. It bundles eight infrastructure services behind one SDK, and every service works on its own or together.
The services
Transactional and marketing email and SMS with 99.9% delivery. Template versioning, batch operations with rate limiting, verification codes, real-time delivery tracking, and compliance tools with GDPR-compliant unsubscribe handling. Over 100M messages processed monthly.
Passwordless login, OAuth with 30+ providers, enterprise SSO via SAML 2.0 and OIDC, directory sync through SCIM 2.0, multi-factor authentication, RBAC, session management, and Auth Guard for threat detection. Token verification in under 10ms. ISO 27001 compliant.
Multi-region file distribution across US, EU, and Asia with 11 nines of durability. Resumable uploads up to 4TB, on-the-fly image processing, signed URLs, object versioning, lifecycle management, and 200+ CDN edge locations. Access control ties directly into Auth.
Define a pricing plan as a bundle of recurring, one-time, and metered prices, then bill a customer with one API call. Subscriptions, trials, proration, invoicing, dunning, tax calculation, and multi-provider routing through Stripe and Paystack across 150+ currencies and 20+ payment methods.
Async infrastructure: durable pipelines with state persistence, queues with priorities and rate limiting, pub/sub event subscriptions, real-time broadcast channels with presence tracking, scheduled tasks, managed Redis clusters, full-text search, and approval gates for human review. Sub-100ms message latency at 500K+ messages per second per queue.
Managed Postgres and SQLite with automated backups, point-in-time recovery, read replicas, connection pooling, auto-failover, and zero-downtime scaling. Conjoin Live streams database changes to your frontend or backend in real time through Live Collections, Live Views, and Change Streams.
Run code, host apps, and power AI agents. Three lifecycle modes in one service: ephemeral executions with sub-500ms cold starts for on-demand code snippets, persistent sessions up to 24 hours for AI agents with writable filesystems and real-time output streaming, and full app hosting with git-based deployments, auto-scaling, scale-to-zero, and custom domains with automatic TLS. The Conjoin SDK is pre-injected into every container, so code can access storage, databases, and messaging without managing credentials.
Context-native AI for enterprises. Feed private data to any supported model while enforcing per-user access controls and redacting PII before it reaches third-party providers. Multi-model inference, declarative workflow orchestration, zero-ETL indexing from Storage, and budget controls per project, team, or user. Supports OpenAI, Anthropic, Gemini, Grok, Mistral, DeepSeek, Cohere, and Llama.
Built for the agentic world
Conjoin is AI-native, and that runs through every layer of the platform. The SDK gives an agent a single integration to every service. The CLI is built on that SDK and uses AI agents of its own, so it carries out what you ask and acts across the stack. AI coding tools like Claude Code, Cursor, and Codex drive the CLI directly, so an agent working in your editor or terminal can set up auth, send messages, query a database, run code, and handle billing. The MCP server exposes the same platform over the Model Context Protocol, so any agent that speaks MCP works against Conjoin natively.
Conjoin runs the agents too. Runtime gives them persistent sessions with writable filesystems and the SDK already injected, and the AI service handles model inference across providers, so an agent can reason, call a model, and act on real infrastructure inside one platform. Every layer is built to be operated by agents and to run them, and that's what makes Conjoin agentic.
Where this is going
The end state for Conjoin is that no developer ever has to think about backend infrastructure again. Not auth, not payments, not messaging, not storage, not queues, not databases. You write your product logic and Conjoin handles every layer beneath it.
Today, building a SaaS product means assembling a half-dozen vendors, each with its own SDK, its own billing, its own failure modes, and its own breaking changes. That tax compounds with every new service you add. Conjoin collapses that entire stack into one integration that scales from a side project to a publicly traded company.
The roadmap is to expand Conjoin's footprint into every region where developers are building, with edge compute, local compliance, and payment methods that work where global providers don't. The goal is to make Conjoin the default infrastructure layer for the next generation of software companies, especially in markets that the incumbents have underserved.
