E2B
Secure cloud sandboxes for running AI-generated code, with Workload Identity, Secrets and Embed.
Best forSecure cloud sandboxes for AI agents to run code
Same ranking as the homepage catalog: name, category and summary matches, weighted by how specific the term is. Esc closes.
21 ai agent infrastructure in our directory, 16 added in November. 17 have a free tier we confirmed, 5 are open source. Pricing on every listing is re-read from the vendor's page on a rolling quarterly cycle.
Editor: Theo · picked by community discussion, not payment
Secure cloud sandboxes for running AI-generated code, with Workload Identity, Secrets and Embed.
Best forSecure cloud sandboxes for AI agents to run code
Cloud headless browser infrastructure for AI agents.
Best forHeadless browsers for web-browsing agents and scraping
Local-first runtime control for AI coding agents. Runs beside Claude Code, Codex and Cursor and holds risky shell commands, secret reads, MCP calls and package installs for your approval.
Best forDevelopers who let Claude Code, Codex or Cursor run commands on their machine
Resolve names in text to Wikidata IDs or your own entity identifiers. Returns supported matches or an unresolved result for downstream agents and knowledge graphs.
Best forDevelopers linking entity mentions to stable IDs
21 tools
Secure cloud sandboxes for running AI-generated code, with Workload Identity, Secrets and Embed.
Cloud headless browser infrastructure for AI agents.
Local-first runtime control for AI coding agents. Runs beside Claude Code, Codex and Cursor and holds risky shell commands, secret reads, MCP calls and package installs for your approval.
Resolve names in text to Wikidata IDs or your own entity identifiers. Returns supported matches or an unresolved result for downstream agents and knowledge graphs.
Give an MCP-compatible AI client workspace file reading, search, patches, command execution and Git inspection through a local coding-tool server.
Amazon data MCP server with 19 tools per the official site (the submission stated 21; the official count of 19 is used here). Freemium: free test quota plus pay-per-call.
A-share market sentiment data subscription with REST, MCP and Agent Skill access; docs site and GitHub repo (lava-mcp) verified.
Reddit sentiment monitoring API — reads public content without OAuth. Paid from $0.50/1k calls. GitHub qovia-mcp release verified.
Real-time Amazon data API for AI agents and developers: turn product pages, search results, seller storefronts, Best Sellers and New Releases into structured JSON through six Amazon-native parsers across 13 documented marketplaces. Freemium with 60 free credits per month, no credit card required.
Real-time Amazon review API that streams product reviews into AI agents and apps: review text, ratings, Verified Purchase status, helpful votes, images and videos in structured JSON, with date and rating filters across major Amazon marketplaces.
Amazon Alexa (formerly Rufus) for Shopping recommendation API: programmatically query what Alexa recommends for thousands of shopping prompts -- recommended ASINs, follow-up questions, contextual replies in structured JSON, with URL-context parameter, multi-turn context and screenshot archival. Freemium with 60 free credits per month, no credit card required.
Voice-of-customer MCP server with 26 tools across 7 platforms (TikTok, Instagram, YouTube, X, Facebook, Pinterest, Trustpilot; Threads and Reddit on request): ask your AI agent about brand sentiment, purchase concerns, competitor share, and risk alerts, with every trend linked back to source posts. Six report modules; 3-minute setup over Streamable HTTP. Freemium: 60 free credits a month, no credit card; Starter $19/month.
A free, read-only MCP server for genomics researchers: look up public genetic variant annotations, gene-disease associations and literature evidence through 11 public tools, with no account and no API key. Case workflows need an API key and a subscription.
Independent, unofficial guide site and API relay for TypeSafe's Jev: ask Jev typed questions (choose one, score on a scale, or true/false) and get an answer with a probability and confidence. Includes a browser grading tool, a Grading API and an MCP server. Free 5 credits a month; one-time credit packs from $9.99. The MCP endpoint is hosted by Jagent, not by TypeSafe AI.
AI API relay for apps, coding tools and agents: one OpenAI-compatible key for GPT, Claude, Gemini, DeepSeek and MiniMax, with per-model prepaid pricing and guides for Codex, Claude Code, Gemini CLI and agent runtimes. No monthly subscription; you top up a balance and pay per use.
Patto is a remote MCP workspace that lets your own AI assistant run brand outreach campaigns — campaign briefs, prospects, drafts and replies in one structured place, with your assistant writing and you releasing.
Continua is a free macOS beta that saves Codex Desktop tasks interrupted by usage limits and automatically resumes them in the same environment once usage is restored. It states it does not bypass, extend or work around usage limits.
Zambo is the verifiable work layer for AI: 153 tools behind one MCP endpoint, cross-AI handover that moves a job between AIs with its record intact, and a verifiable AER-1 receipt for every job. Free tier included.
APIMODELS is a unified AI API for developers — 170+ image, video, language and audio models behind one key, pay as you go, with prices below official list rates and automatic failover.
aiFetchly is an open-source, local-first desktop AI agent for business automation - research, document processing, customer communication, scheduled workflows, plugins and business knowledge, with your data kept on your own computer.
Reqode gives AI coding agents product and architecture context so they can build a whole feature and check it against the spec, with branch-aware rules delivered over MCP to Codex, Cursor and Claude Code. Solo is free; Cloud is $19/user/month. No independent third-party hands-on reviews yet.
No match.
By community discussion in our latest sentiment snapshot, the most-discussed ai agent infrastructure are E2B, Browserbase, HOL Guard, resolve() by WordLift. Each listing links to checked pricing and real user quotes.
Yes. 17 of the 21 ai agent infrastructure we list have a free tier we confirmed on the vendor's site, including E2B, Browserbase, HOL Guard, resolve() by WordLift.
E2B, HOL Guard, Coding Tools MCP, Folklore Clinical Variant Interpretation MCP, aiFetchly.
APIMODELS, from $0.01/month (checked Nov 16, 2026; confirm on the official pricing page).
Order follows real community discussion and our rolling pricing re-checks. Rankings are never sold; sponsored placements are labeled and shown separately.
New tools and pricing changes from our rolling re-checks. Weekly picks are coming soon; one click to unsubscribe.