I built a backend so simple that it fits in a YAML file
The open source LLM router that connects your AI agents and harnesses to any provider in seconds. Subscriptions, pay-per-token, local models, and custom providers.
I built a backend so simple that it fits in a YAML file launched on June 1, 2024 with 102 votes, #577 of 4,672 launches that month and more than 90% of that year's launches. The site was still online in September 2026, 2 years on.
In plain words
Manifest.build is a backend system designed for developers who want minimal infrastructure complexity. Instead of traditional database and server setups, the entire backend configuration fits within a single YAML file. It's aimed at builders prototyping applications or running simple projects without elaborate backend architecture. The approach eliminates typical backend complexity while maintaining core functionality needed for small to medium applications.
written from the facts on this page · September 2026
From the sources
Stop overpaying for AI requests. Manifest routes every OpenClaw request to the cheapest model that can handle it. No prompt stored, no quality loss. Most users cut their bill by up to 70%.
Manifest is the only LLM Gateway that autofixes your bad requests on the fly. Manifest is an open source LLM Gateway that helps builders and teams create reliable agents and workflows, and manage their inference consumption. Connect your subscription plans, BYOK providers and local models with your harness in seconds. Customize your routing and maximize your plans. Manage your inference consumption and settings, and track every request and token through a beautiful interface. Don't let avoidable client errors harm your users' experience and your revenue. Help them connect smoothly to models. Ensure zero downtime by fixing issues on the fly with Manifest Autofix. Our agent repairs the…from manifest.build
Does a similar job
all alternatives →
InsForgeMar 2026 · insforge.dev · ▲622We built InsForge because we wanted to use pure prompts to build a production-grade app faster. In practice, AI isn’t reliable enough when generating code or configuring backend pieces like schema, RLS, auth, or functions. Postgres MCP and Supabase MCP are a good starting point, but they felt more like API wrappers for agents than something designed to guide AI safely through backend tasks. So we built our own MCP server on top of Postgres and added a set of context-engineering tools to make prompt-driven workflows more predictable. It gradually turned into a Postgres-based BaaS. Key…
TrueFoundry AI GatewayDec 2025 · truefoundry.com · ▲315Many teams connecting LLMs to external tools eventually encounter the same architectural issue: as more tools and agents are added, the integration pattern becomes an N×M mesh of direct connections. Each agent implements its own auth, retries, rate limiting, and logging; each tool needs credentials distributed to multiple places and observability becomes fragmented. We built LLM gateway with this goal to provide a single place to manage authentication, authorization, routing, and observability for MCP servers, with a path toward a more general agent-gateway architecture in the future. The…
SimplAI2025 · simplai.ai · ▲589Hey HN, I've been building SimplAI for the past several months — it's a platform for building, testing, and deploying LLM-powered agents and multi-step workflows. The problem I kept running into: spinning up an AI agent pipeline means stitching together prompt management, tool calling, memory, evals, and deployment — often from scratch every time. SimplAI tries to be the layer that handles all of that so you can focus on what your agent actually does. What it does: - Visual + code-first workflow builder for chaining LLM calls, tools, and APIs - Built-in prompt versioning and A/B testing…
Arch2024 · github.com · ▲314Hey HN — I’m Adil from Katanemo (with Salman, Shuguang, and Meiyu) We previously shared an early version of this project as ArchGW. Based on customer feedback, the scope expanded from “LLM routing and model access” into something broader: delivery infrastructure for agentic applications. We renamed it to Plano and reworked the architecture accordingly. The problem On-the-ground AI practitioners will tell you that calling an LLM is not the hard part. The really hard part is delivering agentic applications to production quickly and reliably, then iterating without rewriting system code every…
More ai this month
the category →
Paintings by AI models: programs run through a simulation of oil paint.
AI · 22d ago · stillwet.art
FastRouter.ai▲364Route requests to the right LLM for cost, latency & quality
AI · 23d ago · fastrouter.ai
Source: Show HN post ↗
Launched alongside, June 2024
the whole month →
La Growth Machine▲1,216Create personalized, multi-channel conversations at scale
AI · 2024 · lagrowthmachine.com
PyjamaHR▲1,132Hiring on autopilot. The AI applicant tracking system (ATS).
Work · 2024 · pyjamahr.com