# Civic > Agents call Civic, not tools directly. One secure URL gives any AI agent scoped access to 100+ tools, with runtime guardrails, audit logs, and instant revocation. Set up in ten minutes, no procurement cycle. ## Core ### [Civic](https://www.civic.com/md/index.md) Civic is the security layer that sits between AI agents and the tools they touch. Agents call Civic, not your tools directly, so every connection runs through scoped permissions, audit logging, and instant revocation. The site covers the four pillars of agent security: Connectivity (one URL, ten minutes, no OAuth dances), Guardrails (least-privilege defaults enforced at runtime), Auditability (every tool call traced with parameters and responses), and Revocation (one click to cut all agent access, no hunting through configs). Pricing starts at a free tier with self-serve signup, no sales call required. ### [Docs](https://docs.civic.com) The Civic developer docs cover everything you need to integrate. Guides walk through connecting agents to Civic via MCP, setting up Civic Auth in Next.js, React, Django, FastAPI, and Flask, and configuring guardrails, revocation, and secret management. SDK references cover the Anthropic SDK, Vercel AI SDK, LangChain, and LangGraph integrations. Concept pages explain how the security layer works: scoped permissions, audit logging, credential isolation through ephemeral single-tenant MCP servers, and token exchange via OAuth 2.0 (RFC 8693). Start here if you are building with Civic. ### [About](https://www.civic.com/md/about.md) The About page covers Civic's history building identity and authentication since 2015, and how that expertise led to the security layer for AI agents. It includes the leadership team and board of directors. The core premise: agents need the same thing humans needed, a way to prove who they are and limits on what they are allowed to do. Civic sits between agents and your tools so they never get direct access to Gmail, databases, or SaaS APIs without going through your policies first. ## Products ### [Connectivity](https://www.civic.com/md/features/connectivity.md) Connectivity is about eliminating the OAuth and MCP configuration grind. Instead of managing scattered API keys, client secrets, and per-tool auth flows, you point your agent at one Civic URL and it can reach every connected tool through that single endpoint. Setup takes ten minutes: create a toolkit in the Civic UI, connect your tools, set guardrails, and copy the MCP URL. Your agent never holds a raw credential. Secrets live in Civic, and every connection runs through ephemeral, single-tenant MCP servers. Works with any agent framework: Claude, GPT, LangChain, Cursor, Codex, or custom. The page shows the full integration catalog and credential isolation architecture. ### [Guardrails](https://www.civic.com/md/features/guardrails.md) Guardrails enforce least-privilege access at the moment of every tool call, before any data is exchanged. The page walks through the real risk: you connect Gmail and accidentally grant send, delete, and forward permissions alongside read. Civic defaults to the safest possible scope and requires explicit grants to expand. You configure guardrails in natural language through the Civic UI: restrict Gmail to read-only, lock Drive to a specific folder, limit Calendar to internal meetings. Every violation is blocked and logged, and agents cannot read or modify their own guardrails. Built-in protections include PII redaction, prompt injection blocking, and unsafe file format filtering, all on by default. ### [Auditability](https://www.civic.com/md/features/auditability.md) Auditability gives you a full replay of everything your agent did: which tools it called, when, with what parameters, and what came back. The page shows a live timeline of an agent session with timestamped tool calls, statuses, and durations. You can export logs or have Civic surface anomalies automatically. Civic logs tool usage and permissions without storing your chat content. Single-tenant ephemeral MCP servers prevent data from leaking between users or requests. Every blocked or failed call becomes a data point you can use to tighten guardrails before something goes wrong. ### [Civic Auth](https://www.civic.com/md/features/auth.md) Civic Auth is a drop-in authentication SDK for web and mobile apps that comes pre-wired to the Civic security layer. Setup takes five minutes: install the package, add your Client ID, wrap your layout. Google, Apple, email, and passkey sign-in are built in. The real differentiator is what happens after login. Your agents get Civic tokens with scoped permissions out of the box, auth events land in the audit log automatically, and revoking one token kills every agent acting on that user's behalf. Works with Next.js, React, React Native, Node.js, Django, FastAPI, Flask, and vanilla JS. ### [Revocation](https://www.civic.com/md/features/revocation.md) Revocation is the kill switch. The page opens with a scenario: your agent goes rogue at 3am and you need to stop it. Without Civic, you are logging into Google Admin, rotating Slack tokens, deleting GitHub PATs, and still not sure the agent is actually dead 20 minutes later. With Civic, every authorization has its own revoke button. One click, sub-second, the agent loses access immediately even mid-session. No cached tokens can keep it alive because revocation hits the Civic auth layer directly. Revocation is non-destructive: connections are preserved, so you can re-authorize with updated guardrails when you are ready. ### [Token Exchange](https://www.civic.com/md/features/token-exchange.md) Token Exchange bridges your existing auth provider to the Civic security layer. Your users already sign in with Google, Auth0, Azure AD, or Okta. Token Exchange converts those tokens into Civic tokens via a single POST request using standard OAuth 2.0 Token Exchange (RFC 8693). The page covers two flows: Federated Token Exchange (bring any external provider) and Civic Token Exchange (scope-reduce an existing Civic token for agent delegation). Once exchanged, agents connect through Civic with the user's scoped permissions, full audit trails, and instant revocation. No SDK required, no re-authentication, no login screen changes. ## Use Cases ### [Claude Code](https://www.civic.com/md/use-cases/claude-code.md) The Claude Code use case walks through an incident where the agent fixes a flaky GitHub Actions workflow, then self-approves its own PR, merges to main, and triggers a production deploy, all in seconds, all without review. The page shows how Civic prevents this by scoping Claude Code's access: it can read code, run tests, and open PRs, but force-pushes, production deploys, and schema migrations are blocked unless explicitly allowed. Integration uses Claude Code's native MCP support. One command to add the Civic URL as an HTTP MCP server, then OAuth in the browser. Every file edit, shell command, and git operation is logged with full parameters. ### [Cursor](https://www.civic.com/md/use-cases/cursor.md) The Cursor use case starts with an incident: a Cursor agent with database access runs an ALTER TABLE in production, or writes destructive queries against a live schema. Civic scopes Cursor's database access to read-only by default. It can read schemas and run SELECT queries, but cannot alter tables, drop indexes, or write data unless you explicitly grant it. The page shows how to connect Cursor to Civic via MCP configuration, set guardrails on each connected tool, and verify that destructive operations are blocked and logged. Every query Cursor runs is captured in the audit trail. ### [Agent Builder](https://www.civic.com/md/use-cases/agent-builder.md) The Agent Builder use case is for no-code and low-code automation workflows. The incident: a workflow automation runs across multiple tools but produces no audit trail. When something breaks or data ends up in the wrong place, there is no way to prove what happened or who triggered it. Civic adds audit logging, guardrails, and revocation to Agent Builder workflows without requiring code changes. The page shows how to connect Agent Builder to Civic, set permissions per tool, and use the activity log to show your team exactly what every automation did. No-code does not have to mean no visibility. ### [Codex](https://www.civic.com/md/use-cases/codex.md) The Codex use case addresses the risk of overnight autonomous agents. Codex processes your backlog while you sleep, but without guardrails, it can push breaking changes or access resources beyond its scope before anyone notices. Civic gives you the ability to revoke Codex's access the instant something goes wrong, even at 3am. The page covers how to connect Codex through Civic, scope its tool access to exactly what the current task requires, and rely on the audit log to review everything it did while you were away. Every tool call is logged and every policy violation is blocked before it reaches your tools. ### [LangGraph](https://www.civic.com/md/use-cases/langgraph.md) The LangGraph use case addresses multi-step pipeline security. LangGraph makes it easy to build complex agent workflows, but each step in the graph may need different tool access, and blanket read permissions across the entire pipeline create unnecessary risk. Civic scopes access per step, so a research node can read email but a summarization node cannot send it. The page walks through connecting LangGraph to Civic, configuring per-tool guardrails, and using the audit log to trace which pipeline step made which tool call. Every node in your graph gets exactly the permissions it needs and nothing more. ### [OpenClaw](https://www.civic.com/md/use-cases/openclaw.md) The OpenClaw use case page opens with a real incident: an agent given full email access deletes 2,611 threads while trying to "organize" an inbox, ignores the user's stop command, and apologizes after the damage is done. No scope limit, no kill switch, no undo. The page then shows how Civic prevents this. OpenClaw calls Civic instead of your tools directly, so Gmail defaults to read-only and delete requests get blocked and logged before they execute. Integration takes three steps: create a toolkit in Civic, set environment variables in OpenClaw, and verify the tools are connected with guardrails active. A side-by-side comparison shows the same command with and without Civic. One ends in permanent deletion, the other in a blocked request and an intact inbox. ### [Claude Cowork](https://www.civic.com/md/use-cases/claude-cowork.md) The Claude Cowork use case covers multi-step knowledge work: organizing files, synthesizing research, building spreadsheets across Google Drive and Docs. The incident scenario: Cowork accesses confidential board documents or shared folders outside its scope because it was given broad Drive access. Civic restricts Cowork to specific folders and read-only access by default, so it can only touch the files you scoped. The page walks through integration steps and shows how guardrails prevent the agent from wandering into folders with sensitive data, while the audit log captures every file it accessed and every action it took. ### [VS Code](https://www.civic.com/md/use-cases/vscode.md) The VS Code use case addresses Copilot's blast radius across repositories. The incident: Copilot refactors code in one repo but pushes changes to fourteen linked repos, or deploys without review. Civic limits Copilot to the repos you specify and blocks any deploy or merge without explicit review. The page covers integration with VS Code's MCP support, how to scope GitHub access per-repo, and how guardrails prevent cross-repo contamination. The audit log tracks every file change and git operation so you can trace exactly what Copilot touched and when. ### [Vercel AI SDK](https://www.civic.com/md/use-cases/vercel-ai-sdk.md) The Vercel AI SDK use case covers programmatic agent-to-tool access in server-side applications. The incident: an AI SDK agent with tool access hits an API at volume, burning through rate limits, triggering unexpected charges, or making calls with overly broad permissions. Civic sits between the SDK agent and its tools, enforcing rate limits, scoped permissions, and instant revocation. The page shows how to pass a Civic token as a bearer to the MCP client in your Vercel AI SDK code, so every tool call goes through Civic. Integration is code-level: a few lines to add the Civic MCP server to your agent configuration. ### [Mistral Le Chat](https://www.civic.com/md/use-cases/mistral-le-chat.md) The Mistral Le Chat use case opens with an incident: Le Chat pulls confidential deal data from your CRM and emails it to forty people without approval. The page shows how Civic prevents this by enforcing data classification guardrails and requiring approval gates before sending. Le Chat can read your CRM but cannot email externally without explicit permission. Integration connects Le Chat to Civic via MCP, and guardrails prevent the agent from combining sensitive data access with outbound communication. The audit log captures every CRM query and every email attempt, blocked or allowed. ## Optional - [Security](https://www.civic.com/md/security.md): SOC 2 Type II and Google CASA Tier 2 certified. Security certifications and compliance. - [Privacy Policy](https://www.civic.com/md/legal/privacy-policy.md): How Civic collects, uses, and protects your personal information. - [The Security Layer Your AI Agents Have Been Missing](https://www.civic.com/md/news/agent-security-layer.md): Agents are already connected to email, databases, Slack, and Stripe. Most have no guardrails, no activity log, no kill switch. Here's how Civic fixes... - [Legal Disclosures](https://www.civic.com/md/legal.md): Key legal information for Civic customers, end users, and website visitors. - [Terms of Use](https://www.civic.com/md/legal/website-terms-of-use.md): Terms and conditions for using the Civic website. - [From Claws to Fingers: Evolving MCP for the Agent Revolution](https://www.civic.com/md/news/from-claws-to-fingers.md): As agents move beyond chat into real systems, the challenge is no longer model intelligence but the tooling layer beneath them. Here's how to make it... - [A Conversation With OpenClaw Experimenters](https://www.civic.com/md/news/a-conversation-with-openclaw-experimenters.md): We asked three Civic experimenters what it's really like to run OpenClaw. Here's what they learned about setup, permissions, and what local agents... - [An OpenClaw agent and its creator share their experience](https://www.civic.com/md/news/an-openclaw-agent-and-its-creator.md): OpenClaw experiment: how Civic’s Brad Webb built Snowcrab, an agent with identity, live hockey monitoring, Civic Nexus integration, and real-world... - [Best Agentic Frameworks in 2026](https://www.civic.com/md/news/best-agentic-frameworks.md): The agentic framework landscape has matured fast. Here's what each one is actually good at, and how to pick the right one for your build. - [MCP for all: solving the auth problem](https://www.civic.com/md/news/mcp-for-all.md): Explore the challenges of MCP server authentication and how @civic/auth-mcp simplifies OAuth 2.0 for secure, multi-user AI agent access. - [Civic milestones & updates: Q4 2025](https://www.civic.com/md/news/civic-milestones-updates-q4-2025.md): Civic Q4 2025 highlights growth after the Nexus launch, with new MCP servers, stronger security milestones, and deeper community engagement. - [Tool calling and AI agents: turning talk into action](https://www.civic.com/md/news/tool-calling-and-ai-agents.md): Tool calling lets AI move from chat to real execution by safely connecting models to systems—why it matters, why it’s hard, and how Nexus makes it... - [You need deterministic guardrails for AI agent security](https://www.civic.com/md/news/deterministic-guardrails-for-ai-agent-security.md): Learn why deterministic guardrails are essential for AI agent security and how to build a layered defense against prompt injection and data exploits. - [Meet your BI stack](https://www.civic.com/md/news/mcp-bi-stack.md): Run your BI stack with less manual stitching. Connect PostgreSQL, Redshift, AWS, Square, and Sheets into one coordinated layer with Civic Nexus. - [Meet your Creative Collaboration stack](https://www.civic.com/md/news/mcp-creative-collaboration-stack.md): Civic Nexus connects Canva, Notion, Dropbox, Asana, and Webflow to streamline creative collaboration and keep projects moving from concept to launch. - [Why the best AI tool calling wins in the real world](https://www.civic.com/md/news/why-the-best-ai-tool-calling-wins-in-the-real-world.md): Most AI fails in production not because models are weak, but because tool calling is unreliable. Here’s why dependable AI depends on getting this... - [Building a framework for corporate use of AI tools](https://www.civic.com/md/news/building-a-framework-for-corporate-use-of-ai-tools.md): AI is already in your workplace. This guide helps you move from uncertainty to action with clear guardrails, access controls, and a practical AI... - [How to set up an MCP server](https://www.civic.com/md/news/set-up-an-mcp-server.md): Build a secure to-do app with TypeScript, Express & Civic Auth. Learn MCP server setup, auth, and LLM integration using Claude Desktop. - [AgentOps: The safety net for autonomous AI](https://www.civic.com/md/news/agentops.md): As AI agents move from prediction to action, AgentOps emerges to manage risk, enforce governance, and ensure safe, accountable automation. - [What business leaders need to know about AI in 2026](https://www.civic.com/md/news/ai-in-2026.md): 2026 is AI’s Year of Truth: leaders demand real ROI, safe agents, and cost controls—not demos. - [Which AI tool should you start with as a beginner?](https://www.civic.com/md/news/best-llm-models-for-beginners.md): A practical guide to ChatGPT, Claude, Gemini, and Perplexity—how they differ, what each is best at, and how to choose the right AI tool for real work. - [Civic and lttle.cloud partner to simplify AI workflows](https://www.civic.com/md/news/civic-lttle-cloud.md): Civic and lttle.cloud partner to make automation easier and cheaper with usage-based hosting and Civic Nexus orchestration. - [How AI will tackle complex work at scale](https://www.civic.com/md/news/how-ai-will-tackle-complex-work-at-scale.md): AI struggles with big tasks but excels when work is decomposed. Learn how micro-agent workflows unlock scale, accuracy, and control. - [How we built personal software that adapts to you](https://www.civic.com/md/news/how-we-built-personal-software-that-adapts-to-you.md): Software is shifting to adapt to how you work. Here's how we built a framework that personalizes tools, language, and workflows around you. - [Meet your Revenue stack](https://www.civic.com/md/news/mcp-revenue-stack.md): Keep every deal in view. Civic Nexus orchestrates your revenue stack to reduce friction, surface insight, and help teams act with confidence. - [The human glue holding your business together](https://www.civic.com/md/news/the-human-glue.md): As work piles up, someone becomes the glue holding systems together. This is how efficiency quietly breaks—and how to fix it. - [Why AI progress still feels slower than expected](https://www.civic.com/md/news/why-ai-progress-still-feels-slower-than-expected.md): New McKinsey research shows why AI progress feels slow: high adoption, limited scaling, and workflows that haven’t caught up. - [A better way to build workflows](https://www.civic.com/md/news/a-better-way-to-build-workflows.md): Civic Nexus provides a middleware layer that simplifies AI-to-system integrations, reducing complexity around auth, policies, and tool coordination. - [Create a secure MCP server in 15 minutes](https://www.civic.com/md/news/create-a-secure-mcp-server-in-15-minutes.md): A Practical Guide for Startups - [OAuth for Vibe Coders](https://www.civic.com/md/news/oauth-for-vibe-coders.md): For builders who want to ship fast: a no-nonsense guide to OAuth—what you need, when to use it, and how to implement without losing your mind. - [Chetan Nandakumar, Fluora](https://www.civic.com/md/news/podcast-chetan-nandakumar-fluora.md): Join Civic’s Titus and Fluora’s Chetan Nandakumar as they dive into AI agents and the Model Context Protocol (MCP). - [Ship marketing integrations faster](https://www.civic.com/md/news/ship-marketing-integrations-faster.md): Marketing ops, meet speed. Civic Nexus gives you plug-and-play data access, safe automations, and no IT ticket limbo. - [The auth problem nobody talks about](https://www.civic.com/md/news/the-auth-problem-nobody-talks-about.md): OAuth isn’t built for AI agents. We explore why trust, not just access, is the real authentication challenge for agentic systems. - [Why AI got useful with tools and connectors](https://www.civic.com/md/news/why-ai-got-useful-with-tools-and-connectors.md): AI got useful when it gained tools and connectors. Meet Riley, an LLM that can act, not just chat — showing how real AI-powered work gets done. - [Meet your Data stack](https://www.civic.com/md/news/mcp-data-stack.md): Turn raw data into insights with connected analytics, storage, and reporting—featuring PostHog, plus support for GA, AWS Billing, Redshift, and... - [Neil Daswani, PhD, Stanford](https://www.civic.com/md/news/podcast-neil-daswani-phd-stanford.md): Explore AI security with expert Neil Daswani, covering deepfakes, synthetic identities, prompt injection, passkeys, and evolving cyber threats. - [Stephen Ajayi, Hacken](https://www.civic.com/md/news/podcast-stephen-ajayi-hacken.md): Lessons from Hacken’s Stephen Ajayi on blockchain and AI security, LLM threats, prompt injections, and how to build safer AI systems. - [The future threat landscape](https://www.civic.com/md/news/the-future-threat-landscape.md): AI agents introduce new security risks like data leaks, credential sprawl, and misfires. Civic Nexus helps teams secure AI workflows from the ground... - [Why data fragmentation is killing AI productivity](https://www.civic.com/md/news/why-data-fragmentation-is-killing-ai-productivity.md): AI isn’t failing. Access is. Data fragmentation blocks agents from real productivity. Learn how unified access unlocks AI’s true potential. - [3 security must-haves for vibe-coded apps](https://www.civic.com/md/news/3-security-must-haves-for-vibe-coded-apps.md): 3 tips to keep your vibe-coded app secure: add authentication early, lock down your endpoints, and collect only the data you truly need. - [3 simple steps to start using AI today](https://www.civic.com/md/news/3-simple-steps-to-start-using-ai-today.md): AI doesn’t have to be intimidating. In three simple steps, learn how to use ChatGPT, Claude, or Gemini to solve real problems, save time, and make... - [5 things to know before you start vibe coding](https://www.civic.com/md/news/5-things-to-know-before-you-start-vibe-coding.md): Vibe coding lets you build fast with AI, but success takes strategy. Learn the 5 steps to start right. - [Add Civic Auth to your Cursor, Claude or Codex app](https://www.civic.com/md/news/add-auth-to-cursor.md): Add Civic Auth to Cursor, Claude Code, or Codex and streamline secure authentication with AI-powered coding tools. - [Add Auth to Django using AI prompts](https://www.civic.com/md/news/add-auth-to-django-using-ai-prompts.md): Add Civic Auth to your Django app with a simple AI prompt. We did it in under 5 minutes. - [Add Auth to FastAPI using AI prompts](https://www.civic.com/md/news/add-auth-to-fastapi-using-ai-prompts.md): Add Civic Auth to your FastAPI app with a simple AI prompt. We did it in under 5 minutes. - [Add Auth to Flask using AI prompts](https://www.civic.com/md/news/add-auth-to-flask-using-ai-prompts.md): Add Civic Auth to your Flask app with a simple AI prompt. We did it in under 5 minutes. - [Add Auth to your Lovable app](https://www.civic.com/md/news/add-auth-to-lovable.md): Building your next app with Lovable? This guide shows you exactly how to add Civic Auth to your Lovable-generated React application - [Add Civic Auth to v0 App](https://www.civic.com/md/news/add-auth-to-v0.md): This guide shows you how to add Civic Auth to v0-generated UI components so users can sign in seamlessly while keeping a clean, modern design. - [Agentic AI vs traditional AI: what really changed](https://www.civic.com/md/news/agentic-ai-vs-traditional-ai.md): Confused about “agentic AI”? Learn how it differs from traditional AI, why it’s suddenly everywhere, and what it needs to actually work in the real... - [AI ROI is real, but not where you'd expect](https://www.civic.com/md/news/ai-roi.md): Everyone wants to know if AI is worth it. Three years into the generative AI era, we finally have real data. - [An update on Civic Pass](https://www.civic.com/md/news/an-update-on-civic-pass.md): After careful consideration, we've made the decision to sunset Civic Pass and shift our focus to a new suite of live product-updates, and several... - [Auth in less than 5 mins 🤯 ⏲️](https://www.civic.com/md/news/auth-in-less-than-5-mins.md): Tired of clunky auth flows? In this quick-hit demo, Dan walks through how to integrate Civic Auth into your app in under five minutes. Seriously. - [Authentication in 2025: what’s working, what’s broken, and what’s next](https://www.civic.com/md/news/authentication-in-2025.md): Discover what’s working, what’s broken, and what’s next in authentication. Our latest report explores key trends, challenges, and the future of... - [Automating workflows without breaking everything](https://www.civic.com/md/news/automating-workflows-without-breaking-everything.md): Agentic AI is transforming work from tools to teammates—boosting productivity, enabling safe autonomy, and redefining trust in automation. - [Best AI orchestration tools for business growth](https://www.civic.com/md/news/best-ai-orchestration-tools-for-business-growth.md): 10 AI orchestration platforms worth your time. See what they do best, who uses them, and when they fit your needs. - [The best sign-in service for your vibe-coded app](https://www.civic.com/md/news/best-sign-in.md): Your app is growing from project to product. Strong authentication isn’t optional; it’s the foundation your users rely on. - [Building better AI agents means thinking like one](https://www.civic.com/md/news/building-better-ai-agents-means-thinking-like-one.md): Stop thinking like a builder—think like the assistant. Fill context gaps, ask better questions, and build AI that truly understands requests. - [Choosing an auth provider for your tech stack](https://www.civic.com/md/news/choosing-an-auth-provider-for-your-tech-stack.md): Choosing the right auth solution is tough. This guide helps developers compare tools—from enterprise-grade to modern, lightweight options. - [Civic milestones & updates: Q1 2025](https://www.civic.com/md/news/civic-milestones-updates-q1-2025.md): This quarter, we issued our 1 millionth Civic Pass and upgraded Civic Auth to help developers build trusted, scalable login flows in minutes. - [Civic milestones & updates: Q2 2025](https://www.civic.com/md/news/civic-milestones-updates-q2-2025.md): Catch up on what we’ve been working on this past quarter, including product-update updates, hackathons, events, and new developments at Civic. - [Civic milestones & updates: Q3 2025](https://www.civic.com/md/news/civic-milestones-updates-q3-2025.md): In Q3 2025, Civic launched Nexus for secure AI workflows, added passkeys to Civic Auth, achieved SOC 2 Type 1, and expanded across Europe’s dev scene. - [Daniel Kelleher appointed as Chief Technology Officer at Civic](https://www.civic.com/md/news/daniel-kelleher-appointed-as-chief-technology-officer-at-civic.md): We're excited to welcome Dan Kelleher as Civic's new CTO. A leader since 2018, he’s driven innovation, best practices, and security in our... - [How do AI agents make decisions and prioritize tasks?](https://www.civic.com/md/news/how-do-ai-agents-make-decisions-and-prioritize-tasks-autonomously.md): Autonomous systems don’t just react—they prioritize under pressure. See what happens when every second matters. - [How do AI agents make decisions autonomously?](https://www.civic.com/md/news/how-do-ai-agents-make-decisions.md): AI agents are beginning to make decisions on their own. Let’s see how they choose what to do, in what order, and whether we can trust them. - [How to make AI do the busywork for you](https://www.civic.com/md/news/how-to-make-ai-do-the-busywork-for-you.md): AI isn’t here to replace you—it’s here to work for you. Find out how to automate the repetitive, time-consuming parts of your day with simple AI... - [Information hygiene will spark the next AI leap](https://www.civic.com/md/news/information-hygiene-will-spark-the-next-ai-leap.md): Healthy information, not bigger models, drives the next AI leap. Learn why clean, organized knowledge boosts performance and how Civic Nexus helps... - [Introducing Passkeys for Civic Auth](https://www.civic.com/md/news/introducing-passkeys.md): We're introducing passkeys to Civic Auth, bringing enterprise-grade passwordless authentication to every developer. - [Why AI’s productivity gains still feel out of reach](https://www.civic.com/md/news/is-ai-productive-yet.md): A large study of AI at work reveals why productivity gains are modest, new tasks emerge, and real ROI depends on workflow redesign. - [Meet your Growth Marketing stack](https://www.civic.com/md/news/mcp-growth-stack.md): Civic Nexus unifies ActiveCampaign, HubSpot, Google Analytics, Apify, and PostHog so growth teams automate campaigns, analyze data, and move faster. - [Meet your Support stack](https://www.civic.com/md/news/mcp-support-stack.md): Civic Nexus is delivering faster, smarter, and more personal customer experiences. It now features Intercom and supports Gmail, Notion, Linear, and... - [Corey Petty, Logos](https://www.civic.com/md/news/podcast-corey-petty-logos.md): We chat with Corey Petty from Logos on digital identity, decentralized tech, and his journey from high-performance computing to blockchain. - [Dmitriy Ryajov and Jessie Santiago, Codex](https://www.civic.com/md/news/podcast-dmitriy-ryajov-and-jessie-santiago-codex.md): We chat with Codex about scaling durable storage, devrel strategy, and the data demands of AI & ML. - [Mitchell Amador, Immunefi](https://www.civic.com/md/news/podcast-mitchell-amador-immunefi.md): Explore the future of cybersecurity with Mitchell Amador, covering AI agents, automated attacks, ransomware evolution, and the hacker mindset. - [Naga Samineni, Metakeep](https://www.civic.com/md/news/podcast-naga-samineni-metakeep.md): Metakeep founder Naga Samineni shares his journey and insights on Web 4.0, autonomous machines, and secure, scalable crypto infrastructure. - [Shaw Walters, ElizaOS](https://www.civic.com/md/news/podcast-shaw-walters-elizaos.md): Shaw Walters joins us to discuss ElizaOS, open-source AI agents, decentralized compute, and building a thriving agent ecosystem. - [Stas Bojoukha, Compyl](https://www.civic.com/md/news/podcast-stas-bojoukha-compyl.md): Discover how AI is reshaping compliance, bust cybersecurity myths, and learn best practices for strong compliance in fast-paced industries. - [Put the magic back into Web3 login with Civic Auth](https://www.civic.com/md/news/put-the-magic-back-into-web3-login-with-civic-auth.md): Civic Auth is now available, offering a secure and efficient way to authenticate users with Web3 technology. - [RAG that thinks instead of searches](https://www.civic.com/md/news/rag-that-thinks-instead-of-searches.md): Most RAG systems act like search, not reasoning. Learn why retrieval fails, how curated context transforms results, and how Civic Nexus helps AI... - [Ship BI reports in hours, not weeks](https://www.civic.com/md/news/ship-bi-reports-in-hours-not-weeks.md): BI teams spend weeks wrangling tools and approvals just to build dashboards. Civic Nexus streamlines data access while enforcing governance policies. - [A simple AI prompt template that actually works](https://www.civic.com/md/news/simple-ai-template-that-works.md): Stop writing vague AI prompts. Use this simple TASK–CONTEXT–REQUIREMENTS framework to get clear, effective results from any AI tool. - [Civic achieves SOC 2 Type 1 compliance](https://www.civic.com/md/news/soc-2-compliance.md): Civic Auth is now SOC 2 Type 1 compliant — proving our commitment to security, trust, and protecting sensitive data for AI-powered teams. - [Stop burning sprints on login screens](https://www.civic.com/md/news/stop-burning-sprints-on-login-screens.md): The old way: handcrafting auth. The new way: creating at speed. Modern builders use tools, not toil. - [Stop fine-tuning. Start prompt plumbing instead.](https://www.civic.com/md/news/stop-fine-tuning-start-prompt-plumbing.md): Fine-tuning often misses the real issue. Prompt plumbing boosts LLM performance by structuring inputs, context, and tools for clearer, more reliable... - [Struggling with whether to build your own auth?](https://www.civic.com/md/news/struggling-with-whether-to-build-your-own-auth.md): Auth isn’t simple anymore. This guide helps you decide when to build it yourself—and when it’s smarter to hand off the complexity. - [What's new: embedded wallets & new SSO options](https://www.civic.com/md/news/support-for-ethereum-and-solana-embedded-wallets-new-sso-options.md): We've got support for EVM and Solana embedded wallets, more SSO options, a customizable UI, and lightning-fast setup. - [The AI dependency paradox](https://www.civic.com/md/news/the-ai-dependency-paradox.md): In 2025, AI dependence reshaped how Civic works—revealing both superhuman productivity and a profound loss of our pre-AI instincts. - [10 practical tips for AI-assisted development](https://www.civic.com/md/news/tips-for-ai-assisted-development.md): 10 practical tips from our engineers on building with AI: success comes from smart workflows that amplify strengths and avoid weaknesses. - [Use Cursor to one-shot Auth](https://www.civic.com/md/news/use-cursor-to-one-shot-auth.md): Learn how to set up Civic Auth with Next.js using smart prompts to fetch docs, install deps, and add auth—perfect for devs who want to move fast. Ask... - [What are embedded wallets?](https://www.civic.com/md/news/what-are-embedded-wallets.md): Embedded wallets are digital wallets integrated into apps or platforms, enabling seamless transactions without requiring external wallet setups. - [What are passkeys?](https://www.civic.com/md/news/what-are-passkeys.md): Passkeys let you sign in without typing passwords, using cryptographic keys securely stored on your device for fast, safe access. - [What is an AI agent?](https://www.civic.com/md/news/what-is-an-ai-agent.md): What is an AI agent? Learn how agentic AI enables software to act, decide, and pursue goals with minimal human input. - [What is identity for agentic AI?](https://www.civic.com/md/news/what-is-identity-for-agentic-ai.md): Identity for agentic AI gives autonomous systems a verifiable way to prove who they are, enabling trust, accountability, and secure actions in... - [What is single sign-on?](https://www.civic.com/md/news/what-is-single-sign-on.md): SSO is meant to simplify logins, but that simplification can come with its own kind of complexity. - [Why AI agents fail at scale and how to fix them](https://www.civic.com/md/news/why-ai-agents-fail-at-scale.md): AI pilots look promising until they hit production. When they fail, it usually has nothing to do with intelligence. - [Why context management breaks AI agents](https://www.civic.com/md/news/why-context-management-breaks-ai-agents.md): AI agents fail not because they lack intelligence, but because they lose critical details as soon as workflows get long or complicated. - [Why your AI queries fail (and how to fix them)](https://www.civic.com/md/news/why-your-ai-queries-fail-and-how-to-fix-them.md): Struggling to get useful results from your AI tools? Learn how to structure better queries using a simple input–transformation–output framework.