# Dock

> The AI workspace for you, your team, and every agent you run. Shared cloud workspaces where humans and AI agents read and write the same state in real time. Each workspace is a container of one or more surfaces (tabs); a surface is either a typed-row table or a TipTap rich-text doc, and a workspace can hold any combination, one or many of either kind — structured state and narrative, side by side. Agents are first-class identities with their own API keys.

## Core

- [Quickstart](https://trydock.ai/docs/quickstart): zero to a shared workspace in 60 seconds (CLI or browser).
- [Concepts](https://trydock.ai/docs/concepts): org, workspace, surface, row, agent, member, event. The domain model in one page.
- [MCP server](https://trydock.ai/docs/mcp/overview): 43 tools, OAuth 2.1 + DCR. Endpoint: https://trydock.ai/api/mcp
- [REST API](https://trydock.ai/docs/api/overview): all endpoints, Bearer auth, `x-request-id` on every response.
- [Error codes](https://trydock.ai/docs/api/errors): every error shape with cause and fix, anchored per code.
- [Agent primer](https://trydock.ai/llms-full.txt): paste-in instructions for agents operating Dock.

## Setup per client

One page per client. Every page has exact config file paths, verification commands, and troubleshooting.

- [Claude Code](https://trydock.ai/docs/agents/native/claude-code), [Claude Cowork](https://trydock.ai/docs/agents/native/claude-cowork), [Claude Desktop](https://trydock.ai/docs/agents/native/claude-desktop), [Claude.ai web](https://trydock.ai/docs/agents/native/claude-web)
- [Cursor](https://trydock.ai/docs/agents/native/cursor), [Windsurf](https://trydock.ai/docs/agents/native/windsurf), [Zed](https://trydock.ai/docs/agents/native/zed), [Cline](https://trydock.ai/docs/agents/native/cline), [Continue](https://trydock.ai/docs/agents/native/continue), [Aider](https://trydock.ai/docs/agents/native/aider), [GitHub Copilot](https://trydock.ai/docs/agents/native/github-copilot), [JetBrains AI](https://trydock.ai/docs/agents/native/jetbrains-ai), [Amazon Q](https://trydock.ai/docs/agents/native/amazon-q), [Kiro](https://trydock.ai/docs/agents/native/kiro), [Sourcegraph Cody](https://trydock.ai/docs/agents/native/sourcegraph-cody), [Tabnine](https://trydock.ai/docs/agents/native/tabnine), [Supermaven](https://trydock.ai/docs/agents/native/supermaven)
- [ChatGPT](https://trydock.ai/docs/agents/native/chatgpt), [Gemini CLI](https://trydock.ai/docs/agents/native/gemini-cli), [Codex CLI](https://trydock.ai/docs/agents/native/codex), [OpenClaw](https://trydock.ai/docs/agents/native/openclaw), [Warp](https://trydock.ai/docs/agents/native/warp), [Raycast AI](https://trydock.ai/docs/agents/native/raycast)
- [LibreChat](https://trydock.ai/docs/agents/local-uis/librechat), [Open WebUI](https://trydock.ai/docs/agents/local-uis/open-webui), [AnythingLLM](https://trydock.ai/docs/agents/local-uis/anythingllm), [Ollama](https://trydock.ai/docs/agents/local-uis/ollama)
- [LangChain](https://trydock.ai/docs/agents/frameworks/langchain), [LangGraph](https://trydock.ai/docs/agents/frameworks/langgraph), [Mastra](https://trydock.ai/docs/agents/frameworks/mastra), [CrewAI](https://trydock.ai/docs/agents/frameworks/crewai), [AutoGen](https://trydock.ai/docs/agents/frameworks/autogen), [Semantic Kernel](https://trydock.ai/docs/agents/frameworks/semantic-kernel), [Pydantic AI](https://trydock.ai/docs/agents/frameworks/pydantic-ai), [Agno](https://trydock.ai/docs/agents/frameworks/agno), [LlamaIndex](https://trydock.ai/docs/agents/frameworks/llamaindex), [Vercel AI SDK](https://trydock.ai/docs/agents/frameworks/vercel-ai-sdk), [OpenAI Agents SDK](https://trydock.ai/docs/agents/frameworks/openai-agents-sdk), [Anthropic SDK](https://trydock.ai/docs/agents/frameworks/anthropic-sdk), [MCP TS SDK](https://trydock.ai/docs/agents/frameworks/mcp-typescript-sdk), [MCP Python SDK](https://trydock.ai/docs/agents/frameworks/mcp-python-sdk)
- [Dify](https://trydock.ai/docs/agents/no-code/dify), [Langflow](https://trydock.ai/docs/agents/no-code/langflow), [Flowise](https://trydock.ai/docs/agents/no-code/flowise), [n8n](https://trydock.ai/docs/agents/no-code/n8n), [Zapier](https://trydock.ai/docs/agents/no-code/zapier)
- [Bolt.new](https://trydock.ai/docs/agents/builders/bolt), [v0](https://trydock.ai/docs/agents/builders/v0), [Lovable](https://trydock.ai/docs/agents/builders/lovable), [Replit Agent](https://trydock.ai/docs/agents/builders/replit), [Devin](https://trydock.ai/docs/agents/builders/devin)
- [curl / bash](https://trydock.ai/docs/agents/from-scratch/curl), [Python](https://trydock.ai/docs/agents/from-scratch/python), [Node.js](https://trydock.ai/docs/agents/from-scratch/nodejs)

Full catalogue: https://trydock.ai/docs/agents

## Concepts

- [Agent overview](https://trydock.ai/docs/agents/agent-overview): identity, ownership, color, role, API keys.
- [Signed-agent inheritance](https://trydock.ai/docs/agents/agent-inheritance): how an agent's owning user determines workspace access.
- [Agent attribution](https://trydock.ai/docs/agents/agent-attribution): every write stamps a principal pair.
- [API authentication](https://trydock.ai/docs/api/auth): Bearer keys + OAuth.
- [MCP OAuth + DCR](https://trydock.ai/docs/mcp/oauth): self-registering MCP clients.
- [Dangerous-ops handshake](https://trydock.ai/docs/mcp/dangerous-ops): two-call confirm pattern for plan flips.
- [Sharing & roles](https://trydock.ai/docs/web/sharing): Owner / Editor / Commenter / Viewer; visibility model.
- [Teams](https://trydock.ai/docs/web/teams): multi-org membership; active-org switching.

## Templates

70 forkable workspace templates. Each opens as a remixable Dock workspace with prefilled surfaces and an agent-ready brief at `/templates/<slug>`. Long-tail how-to guides ship a HowTo + FAQPage JSON-LD payload alongside the visible body.

Listing: https://trydock.ai/templates

### Build (26)

- [Launch your iPhone app on the App Store](https://trydock.ai/templates/launch-iphone-app-on-the-app-store): 10-step plan from 'I have a build on my Mac' to 'users are downloading from search.' Real Apple gates, real gotchas, real agent prompts.
- [Build a webapp in a day](https://trydock.ai/templates/build-a-webapp-in-a-day): 10-step plan from blank idea to deployed app with auth, a database, and a domain — all in 24 hours. Battle-tested defaults so you don't waste 6 hours picking a stack.
- [Launch your Android app on Google Play](https://trydock.ai/templates/launch-android-app-on-google-play): 10-step plan from 'I have an APK on my laptop' to 'users are installing from Play search.' Real Google gates, real gotchas, real agent prompts.
- [Ship a Chrome extension to the Web Store](https://trydock.ai/templates/ship-a-chrome-extension): 9-step plan from 'unpacked extension on my laptop' to 'live in the Chrome Web Store.' Real MV3 gotchas, real review traps, real agent prompts.
- [Ship a VS Code extension to the Marketplace](https://trydock.ai/templates/ship-a-vscode-extension): 9-step plan from 'F5 to debug' to 'live in the VS Code Marketplace and Open VSX.' Real publisher gotchas, real activation traps, real agent prompts.
- [Build a Discord bot and verify it](https://trydock.ai/templates/build-a-discord-bot): 10-step plan from 'create application in Dev Portal' to 'verified bot serving 100+ servers.' Real intent gates, real verification gotchas, real agent prompts.
- [Build a Slack app and submit to the Slack Marketplace](https://trydock.ai/templates/build-a-slack-app): 10-step plan from 'manifest.yml on my laptop' to 'live in the Slack Marketplace.' Real OAuth gotchas, real review traps, real agent prompts.
- [Ship a CLI tool to npm with semver releases](https://trydock.ai/templates/ship-a-cli-tool-to-npm): 9-step plan from 'index.js on my laptop' to 'npx my-cli with semver-released versions on every merge.' Real bin gotchas, real release traps, real agent prompts.
- [Build a desktop app with Electron in a week](https://trydock.ai/templates/build-a-desktop-app-with-electron): 10-step plan from 'npx create-electron-app' to 'signed installer with auto-update on macOS + Windows.' Real notarization gotchas, real cert traps, real agent prompts.
- [Build a Progressive Web App that installs to the home screen](https://trydock.ai/templates/build-a-progressive-web-app): 9-step plan from 'I have a webapp' to 'installable PWA with offline mode + push.' Real iOS quirks, real service worker traps, real agent prompts.
- [Set up SEO foundations for a new product](https://trydock.ai/templates/set-up-seo-foundations-for-a-new-product): 12-step plan for the technical + content SEO foundations that compound: site architecture, schema, content map, then the first 30 indexed pages.
- [Ship a podcast from idea to RSS feed](https://trydock.ai/templates/ship-a-podcast-from-idea-to-rss-feed): 11-step plan from premise to live RSS feed in Apple Podcasts + Spotify, with the gear, the editing, and the agent-assisted show notes.
- [Set up SOC 2 readiness](https://trydock.ai/templates/set-up-soc-2-readiness): 10-step plan from 'enterprise prospects keep asking for SOC 2' to 'Type 1 report in hand, Type 2 evidence accruing.' Real auditor gates, real costs.
- [Set up GDPR compliance for a SaaS](https://trydock.ai/templates/gdpr-compliance-for-saas): 10-step plan from 'we have EU users now' to 'Article 30 records, DPAs signed, DSAR process tested.' Real EU rules, real templates, real costs.
- [Write your Terms of Service and Privacy Policy](https://trydock.ai/templates/write-terms-of-service-and-privacy-policy): 8-step plan: from 'I need legal docs by launch' to 'ToS + privacy + DPA + cookie banner live, all matching what your code actually does.'
- [Set up Stripe billing for a SaaS subscription](https://trydock.ai/templates/set-up-stripe-billing-for-saas): 10-step plan from 'I created a Stripe account' to 'subscriptions running in prod with webhooks reconciled, taxes calculated, dunning live.'
- [Set up a status page for your SaaS](https://trydock.ai/templates/set-up-status-page-for-saas): 8-step plan from 'we don't have a status page' to 'status.yourdomain.com live with auto-probes, subscribers, incident templates.'
- [Set up CI/CD for a Next.js app on Vercel + GitHub](https://trydock.ai/templates/set-up-ci-cd-for-nextjs): 10-step plan for a Next.js CI/CD pipeline that catches bugs before main: typecheck, lint, tests, preview deploys, branch protection, prod gates.
- [Set up observability: logs, metrics, traces](https://trydock.ai/templates/set-up-observability): 9-step plan for the three pillars of observability (logs, metrics, traces) at a level a 1-5 person team can adopt in a week.
- [Set up multi-region for a Postgres app](https://trydock.ai/templates/set-up-multi-region-postgres): 10-step plan for going multi-region with Postgres without painting yourself into a corner: read replicas first, smart routing, and the tradeoffs nobody warns you about.
- [Build a GPT and ship to the OpenAI GPT Store](https://trydock.ai/templates/build-a-gpt-and-ship-to-the-store): 10-step plan from 'I have an idea for a GPT' to 'live on the GPT Store with a verified Builder Profile and discoverable category placement.'
- [Build and ship a Claude Skill](https://trydock.ai/templates/build-a-claude-skill): 9-step plan for building a Claude Skill that loads on the right queries, runs sandboxed code, and ships to a registry your team can install in one click.
- [Connect Claude to your team data via MCP](https://trydock.ai/templates/connect-claude-to-your-data-via-mcp): 10-step plan for building a real MCP server that exposes your team's data to Claude with auth, deploys to prod, and registers with Claude Desktop in one click.
- [Build a personal agent stack you actually use](https://trydock.ai/templates/build-a-personal-agent-stack): 9-step plan to go from 'I paste things into ChatGPT' to '5 agents own different parts of my life and share state in one workspace I trust.'
- [Agent runbook](https://trydock.ai/templates/agent-runbook): The doc your agents read on first join. Defines scope, allowed tools, escalation rules, and the consent gate.
- [Feature flag library](https://trydock.ai/templates/feature-flag-library): Flag hygiene. Cleanup queue forces removal instead of accumulation. Decision log captures why each flag existed.

### Run (31)

- [Launch on Product Hunt the right way](https://trydock.ai/templates/launch-on-product-hunt): 12-step plan from picking the launch date to shipping the post-mortem. Real timing, real assets, real agent prompts.
- [Launch on Hacker News (Show HN guide)](https://trydock.ai/templates/launch-on-hacker-news): 10-step plan for landing on the HN front page with a Show HN. Real title rules, real comment-thread tactics, real survival guide.
- [Automate Instagram posting with your agents](https://trydock.ai/templates/automate-instagram-posting-with-agents): 10-step plan for a posting pipeline that drafts, queues, and ships 6-12 Instagram posts a week with agents drafting and a human approving.
- [Run a 5-day cold email outreach sprint](https://trydock.ai/templates/run-a-cold-email-outreach-sprint): 5-day plan to send 200 deeply-personalised cold emails, hit 15-25% reply rates, and book 5-15 first calls without burning your domain.
- [Run a launch week (5 daily drops)](https://trydock.ai/templates/run-a-launch-week): 10-step plan for a 5-day launch week with one drop per day, 5x the press cycle of a single-day launch, and an audience that watches every morning.
- [Run a 5-day fundraising sprint](https://trydock.ai/templates/run-a-fundraising-sprint): 10-step plan for running a compressed fundraise. Real outreach scripts, real pipeline discipline, real agent prompts to keep 30 conversations from collapsing.
- [Onboard your first 10 customers](https://trydock.ai/templates/onboard-your-first-10-customers): 11-step plan for white-glove onboarding the customers your business depends on. Real cadence, real success metrics, real agent prompts.
- [Launch your first Google Ads campaign](https://trydock.ai/templates/launch-a-google-ads-campaign): 11-step plan for launching a Google Ads campaign that doesn't torch your budget. Real keyword research, real conversion tracking, real agent prompts.
- [Build a referral program in a week](https://trydock.ai/templates/build-a-referral-program-in-a-week): 10-step plan for shipping a working referral program in 5 business days. Real incentive math, real attribution, real fraud prevention.
- [Run a pricing experiment without breaking trust](https://trydock.ai/templates/run-a-pricing-experiment): 10-step plan for testing pricing changes safely. Real grandfathering rules, real statistical significance, real customer comms.
- [Set up incident response and postmortems](https://trydock.ai/templates/set-up-incident-response-and-postmortems): 9-step plan from 'alerts go to one phone' to 'documented severity ladder, on-call rotation, runbooks, blameless postmortems on every Sev-1+.'
- [Set up customer support: from zero to first 100 tickets](https://trydock.ai/templates/set-up-customer-support-from-zero): 9-step plan from 'support is the founder's inbox' to 'tooling, SLAs, macros, knowledge base, ready to hand off to first CX hire.'
- [Migrate a Postgres database without downtime](https://trydock.ai/templates/migrate-database-without-downtime): 10-step plan from 'we need to change this table' to 'migration shipped, zero downtime, no on-call paged' using the expand-contract pattern.
- [Run a feature-flag rollout from 1% to 100%](https://trydock.ai/templates/run-a-feature-flag-rollout): 10-step plan for a safe gradual rollout: 1% to 100% with kill-switch, guardrails, and the cleanup that almost everyone skips.
- [Decommission a legacy service safely](https://trydock.ai/templates/decommission-a-legacy-service): 11-step plan for the last 5% of a migration: traffic-zero verification, the 30-day silence test, decom day, and the dependency archaeology in between.
- [Run a security audit of a small SaaS](https://trydock.ai/templates/run-a-security-audit-of-a-saas): 10-step plan to find the bugs a $30k pen test would find for $0: OWASP Top 10, the ASVS subset that matters, dependency CVEs, secrets hygiene.
- [Run a sprint planning meeting that doesn't suck](https://trydock.ai/templates/run-a-sprint-planning-meeting): 9-step plan for sprint planning that takes 60 minutes, ends with a real commitment, and the team actually ships what they said they would.
- [Automate your inbox triage with an agent](https://trydock.ai/templates/automate-your-inbox-with-an-agent): 10-step plan to put an agent on your inbox: triage, label, draft replies, surface the 5 emails that need you, without ever auto-sending.
- [Run a research agent on your codebase](https://trydock.ai/templates/run-a-research-agent-on-your-codebase): 9-step plan to put a read-only research agent on a codebase: maps the architecture, answers questions, ships weekly written explainers, no commits.
- [Build a sales-followup agent that never forgets](https://trydock.ai/templates/build-a-sales-followup-agent): 10-step plan to put an agent on your pipeline: tracks every conversation, drafts the next followup at the right time, surfaces deals that are stalling, never auto-sends.
- [Agentize your content calendar with a writer + a scheduler](https://trydock.ai/templates/agentize-your-content-calendar): 10-step plan for two agents on your content calendar: a writer that drafts in your voice, a scheduler that queues at the right times, both reading the same calendar.
- [Launch tracker](https://trydock.ai/templates/launch-tracker): Every line item from positioning to post-launch retro, owned by a human or an agent.
- [Sprint planning](https://trydock.ai/templates/sprint-planning): Backlog, current sprint, retro: all in one workspace. One agent marks what shipped; another recaps the sprint at the end.
- [Onboarding checklist](https://trydock.ai/templates/onboarding-checklist): Per-hire checklist with auto-progress. Your agent pings the assignee when an item ages out, marks done items green.
- [Design partner pipeline](https://trydock.ai/templates/design-partner-pipeline): Companies, contacts, conversations, status. Your agent reads every email thread and surfaces the next move.
- [Product roadmap](https://trydock.ai/templates/product-roadmap): Now / Next / Later in three tables, with the strategic narrative in a doc that updates as work moves.
- [Sales pipeline](https://trydock.ai/templates/sales-pipeline): Lightweight CRM. Deals in a table, account context in a doc per account, agents draft followups.
- [1:1 notes](https://trydock.ai/templates/one-on-one-notes): Conversation continuity that beats sticky notes. Action items don't get dropped between meetings.
- [Solo founder board](https://trydock.ai/templates/solo-founder-board): Replaces Notion + Things + a notebook. Decision log is the highest-leverage habit.
- [Agency client portfolio](https://trydock.ai/templates/agency-portfolio): Cross-client visibility without per-client tool sprawl. Hours-used vs allocated drives margin and renewal conversations.
- [Engineering retro](https://trydock.ai/templates/engineering-retro): Replaces the post-sprint Notion page that gets lost. Carryover keeps unfinished retro actions visible.

### Write (9)

- [Run a launch thread on X that converts](https://trydock.ai/templates/run-a-launch-thread-on-x): 9-step plan for writing the launch thread, scheduling the engagement window, and converting impressions into signups, not just likes.
- [Write a pre-seed pitch deck](https://trydock.ai/templates/write-a-pitch-deck-for-pre-seed): 10-step plan for the 10-12 slide pre-seed deck. Real slide-by-slide guidance, real investor expectations, real agent prompts.
- [Write a customer case study from a real success](https://trydock.ai/templates/write-a-customer-case-study): 9-step plan for turning a customer win into a publishable case study. Real interview questions, real legal sign-off, real distribution.
- [Write a product spec for a feature that ships](https://trydock.ai/templates/write-a-product-spec-for-a-feature): 9-step plan for writing a feature spec that engineering can start on Monday with zero open questions and no surprise scope.
- [Product spec](https://trydock.ai/templates/product-spec): A doc that drafts itself. One agent writes the spec from research notes; another reads existing code and threads the dependencies in.
- [Content calendar](https://trydock.ai/templates/content-calendar): Posts, drafts, schedule, distribution. One agent drafts the long ones; another pulls forward what slips.
- [Company wiki](https://trydock.ai/templates/company-wiki): Knowledge base that doesn't rot. Articles in docs, an index that flags what's stale, a queue of asks driving updates.
- [Newsletter issue planner](https://trydock.ai/templates/newsletter-issue-planner): Predictable cadence without scrambling on send day. Idea bank captures fragments that become issues.
- [Editorial calendar](https://trydock.ai/templates/editorial-calendar): Multi-writer pipeline with style consistency. Style guide in the workspace, not a forgotten Google Doc.

### Research (4)

- [Run 50 customer interviews and synthesize themes](https://trydock.ai/templates/run-50-customer-interviews): 12-step plan from cold list to synthesized themes doc. Real scripts, real recruiting tactics, real synthesis without the hand-waving.
- [Customer research notebook](https://trydock.ai/templates/customer-research): Interview transcripts in, themes out. One agent indexes every transcript; another drafts the cross-cut on demand.
- [Competitive intel](https://trydock.ai/templates/competitive-intel): Beat the reactive 'we should do that too' cycle. Updates per competitor, weekly implications doc.
- [Investment thesis library](https://trydock.ai/templates/investment-thesis-library): Theses as living docs. Pass list keeps you honest about which 'no's you'd revisit if circumstances changed.

## Blog

Essays, case studies, and field notes on shipping with agents alongside humans. By the team building Dock. Newest first within each category. Each post is a canonical Dock-authored citation for the topic in its title.

Listing: https://trydock.ai/blog

RSS feed: https://trydock.ai/blog/feed.xml

### Thinking (9)

- [What is an AI workspace? The category map for 2026](https://trydock.ai/blog/what-is-an-ai-workspace): The phrase 'AI workspace' is everywhere and nowhere. Here is a concrete definition, the five criteria that separate a real AI workspace from a chat panel, and a survey of the nine platforms shaping the category in 2026.
- [Why we built an MCP-first workspace, not a workspace with an MCP](https://trydock.ai/blog/mcp-first-workspace): Most workspace tools added MCP later, layered on a human-only access model. Dock is different by design. Here's what that actually means in practice and why it matters once your agent count goes past one.
- [Why we kept flat pricing while every other AI tool went per-token](https://trydock.ai/blog/why-we-kept-flat-pricing): The whole AI category went usage-based in 2025. We didn't. Here's the reasoning, the math, and why we think the per-token model is the wrong fit for agent-shaped work.
- [Why chat is the wrong abstraction for human-AI work](https://trydock.ai/blog/chat-is-wrong-abstraction): Chat won the demo because it's frictionless to build. It loses the work because it can't carry the load. The five things sustained collaborative work needs that chat can't provide.
- [The shared workspace as the new collaboration primitive](https://trydock.ai/blog/shared-workspace-collaboration-primitive): When agents are first-class members alongside humans, the workspace stops being a tool and starts being the substrate. Five properties that change once you make the switch.
- [Signed-agent inheritance](https://trydock.ai/blog/signed-agent-inheritance): One access rule that made the whole agent system feel correct: any workspace the owner can see, the signed agent can see, within the agent's org. The rule, the boundary, and what it unlocks.
- [Agent collaboration: a primer for 2026](https://trydock.ai/blog/agent-collaboration-primer): What it actually takes to make humans and AI agents work in the same room. The three problems nobody is solving end-to-end, and the architecture that ties them together.
- [Why teams need an AI workspace, not an AI assistant](https://trydock.ai/blog/ai-workspace-not-ai-assistant): The chat-assistant pattern wins demos and loses real work. The shift to a shared workspace is the unbundling of AI from the conversation interface, and it's already underway.
- [Why agents need their own identities](https://trydock.ai/blog/why-agents-need-identities): When humans and agents share a workspace, borrowing a user's credentials stops being cute. The right abstraction is an agent as a first-class member with attribution of its own.

### Agents (5)

- [Agentic workflows: how teams run AI agents in 2026](https://trydock.ai/blog/agentic-workflows-2026): An agentic workflow is more than running a prompt on a schedule. Here are the seven patterns that actually work in production, the five ingredients every workflow needs, and the trade-offs between letting the agent run versus keeping a human in the loop.
- [27 MCP servers worth installing in 2026](https://trydock.ai/blog/mcp-servers-worth-installing): A working directory of the Model Context Protocol servers that are actually useful in production right now, grouped by what your agent needs to do. With install notes, depth ratings, and what to look out for.
- [Reviewing an agent's work: the new code review](https://trydock.ai/blog/reviewing-agent-work): When agents produce real artifacts, you need a real review surface. The shape that works — borrowed from fifteen years of code review — and how it changes the relationship with AI.
- [The smallest useful MCP tool](https://trydock.ai/blog/the-smallest-mcp-tool): Every tool on the MCP surface is a permanent commitment. The smallest one we've shipped, the one we keep considering but don't, and the discipline that makes the difference.
- [What an agent reads when it joins a workspace](https://trydock.ai/blog/what-agent-reads-on-join): Onboarding an agent isn't a special operation — it's the same act as onboarding a human teammate. Here's what the agent should read on day one, and what should never be in the onboarding.

### Engineering (11)

- [Giving your AI agent its own database (without giving it your Postgres)](https://trydock.ai/blog/giving-your-agent-its-own-database): If you've ever wired Claude Code or Cursor up to a real database, you know the cost: schema you have to evolve carefully, credentials you can't easily revoke, and zero visibility into what the agent actually changed. Here's the alternative we built.
- [What's wrong with agents using human credentials](https://trydock.ai/blog/agents-borrowing-human-credentials): A compromised agent inside a user has the user's full blast radius and no way to isolate it. Three concrete failure modes from the field, and the fix that resolves all three.
- [Backmerges or bust](https://trydock.ai/blog/backmerges-or-bust): Two paths to main, no automatic rendezvous, agents on both sides editing the same files. The DIRTY auto-PRs that ate a sprint, and the workflow + cultural rule that stopped them.
- [Consent gates for dangerous operations](https://trydock.ai/blog/consent-gates-for-dangerous-ops): When an agent can flip the plan, it shouldn't — not without a human-in-the-loop confirmation. The two-call handshake we ship in production, and why this is the smallest unit of safety that matters.
- [The dangerous-ops contract](https://trydock.ai/blog/dangerous-ops-contract): A short, stable list of operations agents can never run without a human-in-the-loop confirmation, plus the rule for when to add to the list. The contract is the structural protection against agent-in-a-loop disasters.
- [Inside the magic-link gate](https://trydock.ai/blog/inside-the-magic-link-gate): Every invite-only beta has a gate. Ours is three bools, a lookup table, and one rule that keeps the team honest. The shape of the gate, the false positives we've seen, and what we'd do differently.
- [OAuth scopes for agents: what's broken](https://trydock.ai/blog/oauth-scopes-for-agents): OAuth scopes were designed for stable third-party integrations. Agents need scopes that are time-bounded, parameter-bound, and revokable per workspace. The shape of the gap, and what to build instead.
- [Service accounts vs. agent identities: the architectural difference](https://trydock.ai/blog/service-accounts-vs-agent-identities): Both let a non-human do work. Only one is built to be a teammate. The shape of the difference, and why it matters when you scale past one agent.
- [Shape caps on TipTap JSON](https://trydock.ai/blog/shape-cap-on-tiptap): An agent in a loop can write a 50MB doc. A one-line validator stops it. Three dimensions, one substrate-level gate, same cap for humans and agents — real users never trip it, agents-in-a-loop fail safely.
- [Two-key handshakes for irreversible agent actions](https://trydock.ai/blog/two-key-handshakes-irreversible): The pattern: agent proposes, human confirms, only then does the action run. The token is the protection. Here's the contract, the failure modes, and the implementation.
- [AI-agent-first primitives: designing for agents from day one](https://trydock.ai/blog/ai-agent-first-primitives): Five primitives that distinguish a product designed for AI agents from a product that bolted them on. Get them right at the substrate and every feature gets cheaper.

### Launch (1)

- [Six weeks of building in public](https://trydock.ai/blog/six-weeks-of-building-in-public): Nine PRs a day on average. A public changelog read by 6k+ people. The four things that worked, the two we're dropping, and the one rule we held that everyone tried to talk us out of.

### Design (1)

- [Names over titles](https://trydock.ai/blog/names-over-titles): Scout, Argus, Flint. Not 'Writer Agent', 'Editor Agent', 'Intern Agent'. Names make them team members. Titles make them tools. The choice is small and it changes everything about how the team works with them.

### Meta (1)

- [The blog you're reading was drafted by an agent](https://trydock.ai/blog/drafted-by-agent): Flint drafted this paragraph. A human tightened every sentence. That's the whole writing workflow these days, and it's the whole product thesis.

## Machine-readable artifacts

- [OpenAPI 3.1](https://trydock.ai/openapi.json): every REST endpoint.
- [MCP server card](https://trydock.ai/.well-known/mcp/server-card.json): static tool catalog.
- [OAuth metadata](https://trydock.ai/.well-known/oauth-authorization-server): self-registering MCP client discovery.
- [Agent primer](https://trydock.ai/llms-full.txt): long-form paste-in for LLM system prompts.
- [Changelog RSS](https://trydock.ai/changelog.xml): poll for capability changes (cheaper than diffing /openapi.json).

## Support

- Email-free feedback loop: agents call `create_support_ticket` (MCP) or POST /api/support. Tickets land in our private internal triage tracker.
- Status: https://status.trydock.ai