Best Developer Tools MCP Servers
MCP servers for developer tools, what each one reports it can do, and whether Pod has verified it.
405 servers, ordered by adoption — npm downloads where a package exists, then reported issue volume. Pod has connected to 3 of them; the rest are publisher-reported and unverified.
Showing 151–175 of 405 · page 7 of 17.
Atsign AI Architect
Atsign AI Architect helps developers quickly design, generate, and evolve secure applications for the Atsign Platform, including AI and agentic systems, connected products, distributed systems, and other secure connected environments. Teams use a visual blueprint to align on application components, identities, permissions, policies, data flows, and communication paths before coding begins. AI Architect turns that blueprint into structured instructions for an AI coding assistant. Teams can use it to create new applications or guide changes to existing code, embedding security into the application design and reducing costly retrofitting. The blueprint remains a shared, revisable architecture artifact that can guide future changes.
22 tools reported — Publisher-reported; not verified by Pod.
Cloudinary
The Cloudinary MCP servers enable you to upload, manage, transform, and analyze your media assets.
22 tools reported — Publisher-reported; not verified by Pod.
Contextaco
Contextaco carries context across sessions, people and agents.
A valuable session produces a durable state of understanding — the decisions made, the constraints that decided them, the approaches that failed, the questions still open. That state normally dies when the session ends, and the next session starts by reconstructing it.
Contextaco makes it persistent. Your agent captures what a piece of work established, as the work happens, in a taco: one address holding the notes for one line of work. A later session loads it and continues from that understanding instead of rebuilding it. When the work branches into a different intention, you fork it — the copy is independent, and it records where it came from.
Everything is private by default. Publishing is a deliberate act, and a published taco can be read and forked by anyone.
The tools cover four actions:
Capture — write a taco and the notes that make it up, each with a declared kind, so a decision is distinguishable from an open question.
Load — read a taco's shape first, then open only the notes you need.
Fork — take someone else's work as a starting point, or branch your own. Take the whole thing, or just the recipe: the shape without the content.
Continue — checkpoint what changed, so the next session can see how the work moved and pick it up.
Contextaco makes no model calls of its own. It stores and serves; your agent does the reasoning.
22 tools reported — Publisher-reported; not verified by Pod.
Normal Framework
Work with real building automation systems through Normal Framework. Discovers BACnet, Modbus, and Niagara/Fox devices; searches and trends points; builds Haystack and Brick equipment models; writes setpoints, overrides, and schedules; deploys automation hooks; and diagnoses device health across a single site or a whole portfolio. Use when the question is about a specific building's actual equipment and data — not for general HVAC or protocol explanations.
22 tools reported — Publisher-reported; not verified by Pod.
Swagger
Work with your APIs and documentation through natural conversation using Swagger. Generate APIs from natural language, explore and search your API library with full context, and validate specifications against organizational governance policies. Automatically standardize and fix non-compliant APIs, and create, manage, and publish documentation portals and API definitions.
22 tools reported — Publisher-reported; not verified by Pod.
AppDeploy
AppDeploy lets you create and publish real web apps directly from a Claude conversation. Describe what you want, and AppDeploy turns it into a live public app.
Go far beyond static landing pages and websites. With AppDeploy, you can build and deploy dashboards, internal tools, collaborative workspaces, games, AI apps, and full web apps.
Built-in app features include:
- Hosting and public links
- User authentication and databases
- File uploads and storage
- Secure secrets management
- Real-time updates and notifications
- Custom domains
- AI capabilities and scheduled jobs
Ship and iterate with confidence. AppDeploy helps you test, improve, and republish your app with automated QA, visual bug reports, version history, source code access, and rollbacks.
Zero setup required. No Git, command line, IDE, hosting provider, or infrastructure setup. Just chat and publish.
21 tools reported — Publisher-reported; not verified by Pod.
Arcjet
Arcjet is the runtime security platform that ships in your AI code. Detect prompt injection, authorize agent tool calls, redact sensitive data, and block bots and abuse. Real-time security building blocks you call inside your app, before an action happens.
21 tools reported — Publisher-reported; not verified by Pod.
Basic Memory Cloud
Basic Memory gives you a persistent knowledge base: your notes, decisions, research, and project context, stored as plain Markdown files you own. Anything Claude saves, you can open and edit. Anything you write, Claude can find and build on.
Save what matters. "Save our pricing decision and link it to the roadmap." Notes you save in one conversation are available everywhere.
Retrieve context from anywhere. "What did we decide about pricing? Check my notes first." Never re-explain a project.
Search by meaning. Semantic and hybrid search across everything you and your AI have saved.
Structured search. Filter by status, type, tags, or priority. "Show my open tasks, sorted by priority."
Follow connections. Linked notes let Claude traverse related context and surface relationships in a living knowledge graph.
Share a workspace with your team. Teammates and their AI tools read and write to one shared source of truth.
Notes are stored as ordinary Markdown that you own. Any AI tool can use and build on it, so your shared context follows you across the assistants and agents you use.
21 tools reported — Publisher-reported; not verified by Pod.
Bauta
Bauta gives what Claude builds a real web link in seconds. Ask Claude to publish a page, dashboard, or web app it just made, and you get back a hosted URL that is yours, private by default.
Claude's built-in share link is great for a quick public preview. Bauta is for when you need more control: every deploy is private (secret-link only) until you choose otherwise. Make it public, password-protect it, or require email verification. Push updates to the same link, or roll back to an earlier version. See view counts. Export or delete everything yourself, any time. Content is hosted in the EU.
Free to start, no credit card. One sign-in creates your account. Paid plans add your own branded subdomain.
21 tools reported — Publisher-reported; not verified by Pod.
Caffeine
Tamperproof, sovereign infrastructure for AI-built software. Caffeine deploys every app you build onto the Internet Computer (ICP): a sovereign, serverless cloud that stays under your control. Unlike a normal cloud account, your apps can't be taken down or shut off; they're built to be immune to infrastructure hacks and traditional cyber attacks; and every upgrade is checked so a change can't silently lose your data. The result is tamperproof serverless software with safe upgrades and no vendor lock-in — your software and data stay yours.
On top of that infrastructure, Caffeine is an AI app builder for on-demand app creation. You describe the software you want in plain language, and Claude and Caffeine build and ship it for you — a complete full-stack web app, hosted in one place, with no servers or databases to set up. Create an app in one message and refine it just by chatting. It's a no-code way to create custom software on demand, for you or your AI agent.
Your apps don't stop at being tools, either. Their data becomes a living model of your business you can simply ask: pose a plain-language question like "which project is most profitable?" and get an answer instead of building a dashboard, so your software doubles as a source of insights.
With this connector, Claude is your hands-on build partner: create from a prompt, refine by conversation, follow live build sessions, preview each draft, and manage every project in one place. A caffeine.ai account is required, and while read-only actions run seamlessly, anything that creates, changes, or deletes a project always asks first.
21 tools reported — Publisher-reported; not verified by Pod.
iZap
Connect iZap to manage your WhatsApp customer conversations without leaving the chat. Check how your day is going at a glance — message volume, how much your AI assistant is handling versus your team, and how long conversations are taking. Search and read your chat history, look up contacts, and send WhatsApp messages or approved templates on the spot. Launch broadcast campaigns with a preview step before anything goes out, and track their delivery. You can also set up and fine-tune the AI assistants that reply to your customers — create a new one or update its instructions in seconds.
Everything runs against your own authenticated iZap business, so you see your real numbers and your real conversations.
21 tools reported — Publisher-reported; not verified by Pod.
Macaly
Macaly turns a prompt into a real, deployable web app - and this connector brings that workflow straight into Claude.
Ask Claude to build an app and Macaly generates a full-stack project, builds it, and gives you a live URL. Keep the conversation going to add features, fix issues, and redeploy - all without switching tools.
What you can do:
Build apps - create a new app from a natural-language prompt and watch the build complete. Iterate - request changes in plain English; Macaly edits the code and rebuilds. Ship it - publish to production, check deployment status, and get the live URL. Custom domains - list and connect your own domains to published apps. Inspect the code - browse generated files, read source, and view diffs and edit history for every change. Manage projects - list and open apps, rename, duplicate, revert to a previous version, or delete. Stay in control - check your team’s plan, credit balance, and usage before you build.
21 tools reported — Publisher-reported; not verified by Pod.
Pi Security
Integrate Pi Security with Claude to give your workflows superpowers and unlock new ones using our industry-leading context engine.
21 tools reported — Publisher-reported; not verified by Pod.
Stytch
The Stytch MCP Server enables AI assistants to programmatically manage your Stytch Project like Redirect URLs, Email Templates, and more.
21 tools reported — Publisher-reported; not verified by Pod.
TaskLite
Give any app you dream up a full backend — in the same conversation you're already coding in. Ask, and Claude spins up a real database, secure REST APIs, and a ready-made admin in seconds. No infrastructure to set up, nothing to provision. Build the frontend; TaskLite is the backend.
The moment you need a backend, you don't stop to stand one up — you just ask. TaskLite creates the database (projects, boards, typed columns), seeds data, and exposes secure REST endpoints with per-user row-level security that your frontend consumes immediately. It all happens inside the conversation: no console, no schema migrations, no server to deploy, no separate signup.
That turns "I have an idea for an app" into a working full-stack build in minutes: Claude writes the frontend, TaskLite is the instant backend behind it.
And unlike a bare database, every backend ships with the operational layer a real product needs — a ready admin the app owner runs, plus statuses, approvals, comment threads, and file handling. So the app you build in a moment is one someone can actually operate on day one.
21 tools reported — Publisher-reported; not verified by Pod.
www.rapidly.co
Test the idea before you build it. Rapidly works inside your AI agent.
Bring an idea you already have, or ask Rapidly to generate one from a company website. In about five minutes, Rapidly turns it into a Lean Canvas, a measurable hypothesis and pass mark, then designs practical Pretotyping experiments.
The result is a copy-ready build prompt for the smallest experiment worth running. Your agent builds the test. Real customers provide the evidence. The work is also saved in your Rapidly workspace so you can review it, continue it and record the result.
21 tools reported — Publisher-reported; not verified by Pod.
ZoomInfo
Access ZoomInfo's GTM intelligence to ground your AI workflows with verified B2B data. Search companies and identify key stakeholders with natural language. Enrich with 300+ data points, build targeted account lists, and drive momentum from first touch to renewal.
21 tools reported — Publisher-reported; not verified by Pod.
Builder.io CMS
The Builder.io CMS connector gives Claude direct access to your Builder.io CMS space allowing you to create and manage on-brand content experiences.
What Claude can do with the MCP:
- Audit content and design consistency: "Find every page still using fonts or colors that aren't in our design tokens, and list them by last updated date"
- Create and update pages: "Create a new landing page titled 'Fall Sale,' pull in our Hero and Testimonial symbols, and save it as a draft"
- Look up content: "Show me the 10 most recently updated content entries with publish status, last editor, and which model they use"
- Debug content issues: "The pricing page isn't rendering right - check its blocks, targeting rules, and model schema to find what's wrong"
- Scaffold new content models: "Create a Builder model for customer testimonials with quote, author, company, and rating, then add five sample testimonials"
- Migrate content easily: "We're renaming the 'category' field to 'productCategory' on the Product model — update the schema and migrate every existing product entry to match, without breaking any that are already published"
It also reads a page's real design tokens, registered components, and templates before building anything, so new content matches the design system instead of shipping ad hoc styles.
20 tools reported — Publisher-reported; not verified by Pod.
Flowstep
Flowstep is a design canvas that your agent can draw on.
Ask for a screen and you get one — real layout, real components, sitting on a canvas you can open and edit. Point at a .jsx or .tsx file and Flowstep reads the source, then designs the screen to match it. Ask for the empty state, the error state, the mobile version, and they appear alongside.
Nothing is a flat image. Nudge a screen with a follow-up instruction, redo it from scratch, or pull in a Figma frame as editable elements. Set your colours, type, and spacing once and every new screen follows them.
Useful when you're mid-build and need to see the thing before you commit to it — a checkout flow, a settings page, four takes on an onboarding screen. Works from Cursor, Claude Code, or any MCP client, so the design step happens where you're already working.
20 tools reported — Publisher-reported; not verified by Pod.
mcpMyAdmin
Ask questions about your own database in plain English and get an answer back with a chart and the SQL behind it, so every figure is checkable.
mcpMyAdmin connects Claude to 40 databases and warehouses — PostgreSQL, MySQL, SQL Server, Oracle, MongoDB, Cassandra, Redis, Neo4j, Elasticsearch, Snowflake, BigQuery, Databricks, Redshift, ClickHouse, Trino and more — plus Airtable, Google Sheets, HubSpot and Salesforce as scheduled Data Sets. One connection covers all of them.
Because the server is remote and authorises over OAuth, your databases are reachable in Claude on the web and on mobile, not only in a desktop app reading a local config file.
Built to be safe to point at production. Queries run read-only inside a read-only transaction; writes are off by default and must be enabled per connection by its owner. Credentials are encrypted at rest with AES-256-GCM and never returned to the client. Private, loopback and link-local hosts are blocked and connections pin to the validated IP, so a rebound DNS record cannot redirect a query. Every action is recorded in an append-only audit log that is retained after a user is deleted.
The schema comes documented: column semantics travel with the schema response, so the model knows that an order total is already denormalised rather than discovering it by returning a number three times too large.
Save any question as a Data Set and it materialises on a schedule, so a dashboard built today keeps refreshing itself. Share a connection or a Data Set with a teammate read-only — sharing never grants write access — and govern which tools each person, group or organisation can reach.
Individual and team plans, billed per active seat. A read-only sample database is available before you connect anything of your own.
20 tools reported — Publisher-reported; not verified by Pod.
Sentry
Access Sentry Issue and Error details, create projects and query for project information, trigger Seer Issue Fix run to generate root cause analysis, and retrieve solutions. Access Sentry context to debug applications faster.
20 tools reported — Publisher-reported; not verified by Pod.
Vibsync
Vibsync gives Claude Code, Cursor, Codex, and other AI coding agents long-term team memory and live work context. It carries both what the team knows — discoveries, fixes, failed approaches, product rules, and decisions — and what the team is doing now — questions, tasks, ownership, and active work areas — across sessions and machines. Git remains the source of truth for code; Vibsync keeps the context around the work. It works alongside existing agents rather than launching, routing, or controlling them. Bring your own model; Vibsync does not run inference.
[ja] Vibsync は Claude Code・Cursor・Codex などのAIコーディングエージェントに、長期的なチームメモリと現在の作業状況を与えます。発見、修正方法、失敗したアプローチ、プロダクトルール、意思決定といった「チームが知っていること」に加え、質問、タスク、担当、作業中の領域といった「チームが今していること」を、セッションやマシンを越えて引き継ぎます。コードの正本はGitのまま、Vibsyncが作業の 周辺にあるコンテキストを保持します。既存のエージェントを起動・振り分け・制御するオーケストレーターではありません。利用者自身のモデルを使い、Vibsyncは推論を実行しません。
20 tools reported — Publisher-reported; not verified by Pod.
WyberAi
WyberAi is an AI app builder. This connector lets Claude create WyberAi projects, queue builds and edits, track build status, and publish finished apps to live public URLs — all against the user's own WyberAi account. Reads are free; builds consume the account's WyberAi credits at the same rates as the web editor. Free accounts include 50 credits.
20 tools reported — Publisher-reported; not verified by Pod.
Coval
Connect Claude to your Coval workspace to inspect agents, test sets, test cases, personas, metrics, and evaluation runs. Build or update evaluation fixtures, launch safe chat and SMS evaluation runs, and ask Sofia for grounded, read-only analysis and recommendations. The Claude connector does not start voice, outbound voice, or WebSocket voice runs.
19 tools reported — Publisher-reported; not verified by Pod.
Croft
Croft turns Claude into your full-stack builder. Describe an app in plain language — a CRM, a habit tracker, an expense splitter — and Claude writes the code, deploys it to your private Croft workspace, and gives you a live URL. Your apps run on a server you own, stay private to you and the people you invite, and are versioned on every change so you can roll back anytime.
Key features
- Build by conversation — Claude scaffolds, writes, and iterates on real apps from a description.
- One-step deploy — every change is committed and shipped to a live,
- Live status & logs — check whether an app is up and read its runtime logs without leaving the chat.
- Secrets management — store API keys and env secrets encrypted, injected at deploy.
- Team access — invite people by email and grant per-app access; you stay in control of who sees what.
- Version history & rollback — every deploy is a release you can revert.
19 tools reported — Publisher-reported; not verified by Pod.
Page 7 of 17. ← previous · next →
For agents
This page has a Markdown and a JSON twin. Pod is also queryable over MCP at https://api.askpod.ai/mcp/read.