Ce que Render a publié en juillet 2026

Render shifts docs to API-like structure for coding agents

Render introduced structural changes to documentation to support coding agents (e.g., Claude Code, Cursor) as a second reader alongside humans. Docs now require stable contracts, explicit inputs, and …

Render explains choosing between cron jobs, workers, and workflows

Render published a decision framework for selecting async primitives (cron jobs, background workers, workflows) based on failure semantics, retry behavior, and cost. The article contrasts Render's ser…

Render enhances preview environments for AI agent-generated PRs

Render introduced enhancements to its preview environments to address verification gaps for AI coding agent pull requests. The updates enable automatic, isolated full-stack previews per PR, including …

Render enhances monorepo deployment with root directories, build filters, and shared config

Render published a guide outlining how to deploy five services from a single monorepo using root directories, build filters, shared packages, environment variable groups, and private networking. The a…

Render publishes guide on integrating LangChain SQL tools with Postgres

Render released a technical guide explaining how to connect AI agents built with LangChain to a SQL database using LangChain's SQL database tools. The guide details setup prerequisites, tool decomposi…

Render explains sandbox isolation for agent-generated code

Render details how sandboxes isolate untrusted agent-generated code from trusted infrastructure, reducing blast radius from accidental or malicious execution. It introduces a three-role architecture—t…

Render enables managed PgBouncer pooling for Postgres to solve fan-out failures

Render introduced managed PgBouncer connection pooling for Postgres on July 1, 2026, addressing a common fan-out failure where 200+ concurrent tasks exhaust Postgres connection limits. The pooling fea…

Render introduces CLI-based Postgres provisioning with render pg create

Render launched `render pg create`, a CLI command enabling programmatic provisioning of managed Postgres instances for coding agents and scripts. The command requires explicit parameters (name, plan, …

Render introduces suspend/resume workflows for AI agents with Postgres state

Render detailed a new suspend/resume pattern for pausing AI agents mid-run for human approval, replacing naive blocking-wait approaches. The solution uses Postgres to persist workflow state durably, e…

Render introduces three-tier memory architecture for AI agents with Postgres, pgvector, and Key Value

Render published a technical guide outlining a three-tier memory system for AI agents, mapping short-term memory to Key Value, semantic memory to Postgres with pgvector, and long-term memory to relati…

Render launches $10k migration credits to attract workloads from other clouds

Render introduced a migration credits program offering up to $10,000 in credits to incentivize moving workloads from other cloud providers. The program includes white-glove migration reviews and aims …

Render adds HIPAA-enabled workspaces and compliance certifications for AI apps

Render introduced HIPAA-enabled workspaces on Scale/Enterprise plans, offering SOC 2 Type 2, ISO 27001, and GDPR DPA compliance with zero-config private networking for AI applications. A 20% fee appli…

Render adds Anthropic and OpenAI support to managed OIDC authentication

Render expanded its managed OpenID Connect (OIDC) authentication to support Anthropic and OpenAI, allowing services to authenticate without storing long-lived API keys. Short-lived tokens are issued a…

Render MCP server adds OAuth for Claude Code, Codex, and Cursor

Render’s MCP server now supports OAuth authentication for AI coding tools Claude Code, Codex, and Cursor, expanding beyond API key-based access. API key auth remains available for CI/CD and other non-…

Render adds trigger_deploy tool to MCP server for service deployments

Render introduced a new `trigger_deploy` tool in its MCP server, allowing users and agents to initiate service deployments manually. This feature supports retrying failed deploys or forcing clean rebu…

Render updates default Bun version to 1.3.14 for new services

Render changed the default Bun runtime version for newly created services to 1.3.14, aligning with the latest Bun release. Existing services retain their original versions to avoid disruptions.

Render launches generally available AWS OIDC authentication for Pro+ workspaces

Render made OpenID Connect (OIDC) authentication for AWS generally available to Pro and higher-tier workspaces. Services can now assume AWS IAM roles at runtime without long-lived credentials, with Re…

Render publishes guide on PostgreSQL performance optimization for web apps

Render published a technical guide outlining PostgreSQL performance optimization techniques for web applications, including EXPLAIN analysis, strategic indexing, connection pooling, and maintenance op…

Render publishes guide to implementing continuous deployment workflows

Render released a detailed guide explaining how to implement continuous deployment (CD) on their platform, distinguishing it from continuous delivery and outlining prerequisites, repository setup, aut…

Pricing change detected for Render

Pricing updated for Render: - Free: allowance changed from 1 service to 1 service, 512 MB RAM, 1 GB storage - Starter: allowance changed from 1 service to 1 service, 1 GB RAM, 1 GB storage - Pro: allo…

Pricing updated for Render: - Pro: allowance changed from 1 service to 2 services - Team: allowance changed from 1 service to 5 services

Render publishes guide on implementing authentication and authorization patterns

Render released a technical guide detailing authentication and authorization implementation patterns, including session-based, token-based (JWT), and third-party authentication strategies. The post pr…

Render publishes guide to hosting PocketBase with persistent storage

Render published a technical guide explaining how to deploy PocketBase—a single-binary SQLite-backed backend—on their platform using persistent disks to avoid data loss on redeploys. The guide covers …

Render launches as a Heroku alternative with enhanced flexibility and pricing

Render positions itself as a superior alternative to Heroku, which transitioned to maintenance mode in February 2026. Render highlights key advantages including longer HTTP response limits, persistent…

Render enhances AI app deployment with Workflows and Blueprint templates

Render introduced Workflows, a new public-beta feature for running multi-step AI agent pipelines and long-running jobs with elastic compute that scales to zero. The platform also promoted its Blueprin…

Render highlights 5 JavaScript/TypeScript apps deployable via one-click templates

Render published a guide featuring five JavaScript/TypeScript applications deployable via one-click templates, emphasizing AI-driven use cases like voice agents, stateful AI agents, MCP servers, codin…

Render adds free connection pooling to Postgres databases

Render introduced connection pooling for paid Postgres databases using PgBouncer at no extra cost, allowing higher concurrent client connections without upgrading instance types. The feature runs on t…

Best cloud platform to run agents: Render’s evaluation framework

Render published a guide outlining how to evaluate cloud platforms for running AI agents, focusing on three execution shapes: persistent loops, scheduled runs, and event-triggered invocations. It maps…