Changelog

Goose crossed 50,000 GitHub stars, raising independent-evidence scoring and lifting its editorial rating.

Integration depth raised to 4 after confirming Gojiberry publishes an MCP server for running prospecting and outreach from Claude, alongside native HubSpot and Pipedrive sync and 15+ enrichment providers. Displayed rating moved 3.8 to 4.1.

Integration depth raised to 5 after confirming Devin publishes an official MCP server (mcp.devin.ai) plus the DeepWiki MCP server and an MCP client marketplace, on top of native GitHub, Slack, Linear, Jira, and VS Code integrations and a full API. Displayed rating moved from 4.4 to 4.7.

Viktor shipped a one-click Supabase integration giving the agent native access to database operations, edge functions, and log inspection.

Make added an MCP Client app (verified and maintained by Make), so the platform now both exposes an MCP server for external AI tools and connects out to external MCP servers as a client.

VoiceFleet, a 24/7 AI voice receptionist for small businesses that answers calls, books appointments, and routes urgent calls in 30+ languages, joins the directory from €99 per month.

Noded, a post-sales customer context graph that auto-builds a live account story from 20+ tools and exposes an MCP server, joins the directory from $100 per seat monthly.

Sub-scores rebalanced on re-audit: IndEvid 5 to 4 (27 G2 + 223 Trustpilot, not 200+ on a primary platform), IntDepth 4 to 5 (official MCP server confirmed in the registry plus native integrations), PriceTrans 3 to 4 (three public tiers). Total unchanged at 3.95.

Restructured from two tiers (Personal, Organization) to four (Basic, Pro, Team, Enterprise), with MCP access now metered in monthly credits (250 on Basic, 2,500 on Pro).

Individual pricing tiers (Free, Essential, Advanced) are now fully public; pricing transparency improved from partial to mostly-public.

Deep Research moved from a fixed monthly report count to usage-based limits (5-hour and weekly windows) on the Gemini 3 model family, and is now available on the free Gemini tier.

Roo Code VS Code extension shut down May 15, 2026 and repo archived; team directs users to ZooCode fork and Cline.

Launched ClinePass, a $9.99/mo subscription bundling open-weight models (promo $4.99 first month).

MCP support now generally available as both an MCP server and client.

MCP support now generally available as both an MCP server and client.

Launched a native MCP server, letting external AI assistants connect into the product.

Launched a native MCP server, letting external AI assistants connect into the product.

On Business and Enterprise, ChatGPT Agent enforces monthly message limits (40 on Plus, 400 on Pro, 40 on Business and Enterprise), and workspaces on flexible pricing pay 30 credits per message beyond that. Excel and Sheets extensions moved to metered usage; PowerPoint generation stays free through August 6, 2026 before moving to the same model.

ChatGPT Business consolidated into a single $20/user/month (annual, $25 monthly) plan combining ChatGPT and Codex, with agent, Codex, and Office extension usage now drawing from an included pool and metered by tokens, plus purchasable credits for usage beyond the pool.

G2 presence established: 29 reviews at 4.7/5 under claimed Spellbook profile. Previously showed 0 reviews.

Pro tier increased from $10/month to $12/month. Deep tier reviews changed from unlimited to 200 per month.

Token-based pricing confirmed at $2.00 per 1 million tokens effective August 1, 2026 for WhatsApp Business Platform users. Free service messages end October 1, 2026.

Editorial rating upgraded from 4.30 to 4.55. IndEvid 4 to 5 (G2 proxy crossed 200-review threshold at 232 reviews). PriceTrans 3 to 4 (token pricing published on Meta developer docs).

Pricing discovered: three public tiers at $1,000/month (warm intros only), $3,000/month (cold outreach + warm intros), and $5,000/month (full service). Previously listed as custom/quote-only.

Editorial rating upgraded from On Our Radar (2.65) to full rating 3.70 after discovering public pricing at /service-plan and verifying strong Product Hunt performance (1,521 upvotes, triple award winner).

SOC 2 Type II and HIPAA certifications confirmed on Workforce Wave security page, resolving the previously documented SOC 2 contradiction.

ISO 42001 certification added to Harvey AI based on homepage security badges.

Modash increased pricing: Essentials $199 to $299/month, Performance $499 to $599/month.

Kiro added Claude Sonnet 5 model support, iOS app in early access, and parallel agents capability for building across large codebases.

Clay launched Audiences product for centralizing first and third-party GTM data sources, and expanded data marketplace from 150+ to 200+ providers.

Sprout Social editorial rating increased from 4.2 to 4.3: PriceTrans upgraded from 3 to 5 based on 3 fully public pricing tiers with feature comparison and 30-day free trial.

Seventh Sense SOC 2 Type II certification confirmed on vendor homepage.

Surfer editorial rating increased from 4.3 to 4.4: PriceTrans upgraded from 3 to 5 based on 4 fully public pricing tiers with feature comparison.

Surfer added new Discovery tier at $49/month (billed yearly), lowering entry point from $99/month. Top tier renamed to Peace of Mind at $299/month.

Luma Agents editorial rating decreased from 4.0 to 3.7: IndEvid corrected from 4 to 3 (G2 unclaimed with 1 review, Trustpilot 58 reviews at 1.5/5, $130M+ funding qualifies for IndEvid 3).

Opus Clip editorial rating decreased from 4.2 to 3.9: IndEvid corrected from 5 to 4 (G2 has only 6 reviews, Trustpilot 448 reviews at 3.9/5 below 4.0 threshold).

Smartly.io editorial rating increased from 4.1 to 4.3: IntDepth upgraded from 4 to 5 based on seven native bidirectional integrations including new Amazon Ads channel.

Smartly launched Synapse, an AI optimization layer with autonomous Planning Agents and Execution Agents powered by context-aware memory for cross-channel campaign orchestration.

Klaviyo restructured pricing: lowest paid Email tier now $45/month (was $20/month), Composer (formerly Marketing Agent) is a separate credit-based add-on at $110/month for 16,000 credits.

Klaviyo AI editorial rating increased from 4.3 to 4.65: AutCap upgraded from 3 to 4 based on Composer autonomous campaign creation and Customer Agent 65% autonomous resolution.

Ploy: AI marketing platform from Webflow co-founder Bryant Chou (YC S26, $27M seed). Autonomous website building, SEO/AEO optimization, visitor identification, and ad campaigns. Free tier; Starter $50/mo; Pro $300/mo.

Added WhatsApp, Google Calendar, and Microsoft 365 to integrations. WhatsApp confirmed as native platform channel alongside calling, SMS, and email.

Autonomous customer support agent with per-resolution pricing from $0.19. USPTO-patented anti-hallucination architecture. ISO 27001 certified. 60-day performance guarantee.

Editorial rating increased from 4.05 to 4.30 after IntDepth moved from 4 to 5 following native MCP server launch.

Pipedrive launched a native MCP server on June 30, 2026, enabling AI assistants to connect directly to CRM data.

G2 review count updated from 489 to 492.

Security certifications cleaned to approved list only. Removed 6 non-approved Workspace feature entries. Added FedRAMP. Normalized HIPAA-eligible to HIPAA.

Google AI Plus entry price dropped from $7.99/mo to $4.99/mo. Starting_price updated from 7 to 5.

Google AI Plus entry price dropped from $7.99/mo to $4.99/mo; starting_price updated from 7 to 5

Crayon added: enterprise AI competitive intelligence platform with 20+ native integrations, official MCP server, SOC 2 Type II and GDPR compliance at custom enterprise pricing. 385 G2 reviews at 4.6 stars.

Klue added: enterprise AI competitive intelligence and win-loss analysis platform with Compete Agent, battlecards, and 12 native integrations including Salesforce, HubSpot, and Gong. 443 G2 reviews at 4.7 stars.

Outmano added: AI-powered competitive intelligence for B2B SaaS teams, tracking competitor pricing, SEO, content, and reviews with weekly digests from $49/mo.

Editorial rating corrected from 4.50 to 4.30. PriceTrans upgraded 3 to 5. IndEvid corrected from 5 to 4 (no standalone review presence).

Claude Fable 5 restored July 1 after export controls lifted June 30. Sonnet 5 launched June 30 with near-Opus performance at $2/$10 per million tokens.

Claude Code G2 reviews increased from 79 to 89 at 4.7/5. GitHub stars updated from 133k to 135k.

Editorial rating corrected from 4.85 to 4.65. PriceTrans upgraded 3 to 5 (pricing fully public). IndEvid corrected from 5 to 4 (no standalone G2 listing or GitHub presence).

Editorial rating corrected from 4.50 to 4.30. PriceTrans upgraded 3 to 5. IndEvid corrected from 5 to 4 (no standalone review presence).

Editorial rating corrected from 4.50 to 4.30. PriceTrans upgraded 3 to 5. IndEvid corrected from 5 to 4 (no standalone review presence).

Microsoft Scout pricing model updated from flat subscription to usage-based Copilot Credits. The $99/user/mo is the M365 E7 prerequisite license; Scout usage billed separately via credits with pricing not yet published.

Editorial rating decreased from 4.20 to 4.10 due to pricing transparency downgrade from public to partial (Scout credit pricing not yet published).

Editorial rating increased from 4.55 to 4.65 due to pricing transparency improvement from partial to fully public.

Botpress repriced Plus tier from $150/mo to $410/mo annual ($490/mo monthly), folded AI usage into all paid plans, expanded Free tier to 3 seats and 3 AI agents.

G2 reviews updated from 4 reviews at 3.4/5 to 5 reviews at 3.5/5. No rating or pricing changes.

Pricing restructured into three separate product tabs: Outreach, Instantly Credits, and VIP. New Light Speed tier at $358/mo. Credits now separate subscription with Growth $47, Supersonic $97, Hyper Credits from $197. New features: AI Workflows, Web Researcher Agent, AI Email Writer Agent, 5 major LLMs with BYOK.

Chloe AI shared pool credit caps now documented: Essentials 10K, Growth 15K, Scale 20K per month. Named customer outcomes added: FollowUp.com (50% more booked appointments), Flatiron School. Additional organizations add-on documented.

AI Writing word limits now specified per tier (5K to 1M words/mo). Workflow limits tiered at 2/5/50/500 per team. Projects (BETA) on all plans. Recommendations (BETA) added. G2 reviews updated to 9,673. Pricing details expanded with daily email send limits.

New ChatGPT ads monitoring feature added across all plans. Starter tier now tracks ChatGPT, Gemini, and Google AI Overviews (previously ChatGPT only). Action Center moved to Enterprise-only. G2 reviews updated to 2,114. NP Digital confirmed as named customer.

G2 profile now has 1 review at 5.0/5 (previously zero). Data training confirmed as no per vendor homepage. Starting price set at EUR 12.99/mo (Pro tier). Free tier includes 50 AI Agent credits. Inro AI update shipped April 2026.

Homepage pricing inconsistency resolved: all tiers now consistent across site. Starter $99, Growth $249, Platform $499, Enterprise custom. No setup fees, no contracts. Security certification corrected: HIPAA BAA is not a third-party certification. SOC 2 Type II remains in progress with Q3 2025 target passed.

Completed full re-audit of 13 agents across 5 categories: OpenHands, Workforce Wave, Influential, Inro, Jacquard, Writesonic, Apollo.io, Close CRM, Instantly AI Sales Agent, Instantly.ai, lemlist, Make.com, and Claude. All listings updated with current pricing, G2 review counts, product changes, run-on sentence fixes, and US English standardization.

Domain migrated from all-hands.dev to openhands.dev. Company rebranded from All Hands AI to OpenHands. GitHub stars updated to 78,508. Self-hosted deployment method added. All empty fields populated.

Claude Design added to Pro plan features. Claude Tag launched for Slack task delegation (Team and Enterprise). Trusted Devices for remote Claude Code sessions added. G2 reviews updated to 355. Fable 5 remains suspended under export control directive.

Make AI Agents graduated from beta to GA. Library of Agents added for ready-made deployable agents. Perplexity AI and DeepSeek AI added as named app integrations. G2 reviews updated to 329. Make Waves 26 conference announced for Prague, October 2026.

Email plan price increased from $31 to $55/mo annual. Email plan volume increased from 5,000 to 50,000 emails/mo. New lemAgent feature added for autonomous prospecting workflows. Intent Signal Agents and Data Enrichment Agents added. G2 reviews updated to 1,709.

Same pricing restructure as Instantly AI Sales Agent. Light Speed tier added. Credits separated into own subscription tiers. Functional full-stack entry point clarified at $94/mo (Outreach + Credits). New AI Workflows and Web Researcher Agent features added.

Full re-audit following vendor submission from Corey Ferreira. Pricing completely restructured: per-editor tiers replaced with usage-based Agencies track and demo-based Businesses track. Old customer references replaced with current case studies (Turo, StubHub International, Trilogy). Voiceflow Core proprietary model launched May 2026. Verified badge granted.

Pipedrive AI affiliate partnership activated. Listing updated with verified pricing and integration data.

Updated the MCP server with buyer decision fields: pricing_transparency, contract_type, data_training, human_in_loop, mcp_compatible, security_certifications, editorial_rating_notes, G2 data, and last_verified_at. AI systems querying the MCP server now get the full structured buyer context for every agent.

Completed full Chrome re-audits across ai-sales-agents (6 agents), ai-research-agents (2 agents), and ai-workflow-agents (partial). All listings verified against live pricing pages, G2 profiles, security trust centers, and privacy policies. Notable rating changes: Helpscout 3.5 to 4.1, Chatbase 3.5 to 3.7, Balto 3.8 to 4.1, Forethought 3.9 to 4.1, SalesMind AI 3.3 to 3.6, SalesLabel 3.1 to 2.7 (On Our Radar), Relevance AI 3.4 to 3.9, ChatGPT Deep Research 4.5 to 4.8, Gemini 4.5 to 4.7.

Added NOVA to AI Coding Agents. Free, open-source terminal coding agent with self-healing execution. Describe your goal, NOVA writes files, auto-fixes errors, resolves dependencies, and handles git. On Our Radar.

Added Elentaria to AI Marketing Agents. GTM operator for B2B that scores 22 channels, builds a 90-day plan, and executes weekly on your stack. Paired with a senior GTM expert. Custom pricing. On Our Radar.

Added Fundraisly to AI Sales Agents. AI fundraising agent that analyzes 300K+ investors, maps warm intro paths from your network, and runs cold outreach to book qualified investor meetings. Custom pricing. On Our Radar.

Added Meta Business Agent to AI Sales Agents. Meta's AI agent for businesses on WhatsApp, Messenger, and Instagram. Answers questions, recommends products, books appointments, and facilitates sales. Free setup with paid tiers coming.

Added Gemini Spark to AI Workflow Agents. Google's 24/7 personal AI agent that runs autonomously in the cloud. Tasks, Schedules, and Skills frameworks. Requires Google AI Ultra at $100/month. US only at launch.

Added Microsoft Scout to AI Workflow Agents. Microsoft's first always-on personal AI agent for M365 with its own governed Entra identity. Bundled with M365 E7 Frontier Suite at $99/user/month. MCP server support for third-party tool connections.

MCP server support launched February 2026. Jules now connects to Linear, Stitch, Neon, Tinybird, Context7, and Supabase through vetted integrations, with more partners planned.

Acquired by Cognition AI in July 2025. Windsurf 2.0 launched April 2026 with an Agent Command Center and embedded Devin integration. Now listed under developer Cognition.

MCP server support added to the v0 API as of March 2026. Teams can configure custom MCP servers and reference them directly within v0 chat sessions.

MCP support added with full OAuth 2.0/2.1 authentication for remote MCP servers. Product has also pivoted from a GitHub PR bot to a dedicated JetBrains coding assistant.

Completed full rewrite of all 25 editorial guides to the new content standard. Every guide now has a 4-paragraph introduction, multi-paragraph section content, evaluation criteria specific to the use case, a decision framework, and 5 substantive FAQs. Covers all 8 agent categories plus buying frameworks, ROI calculation, and foundational explainers.

Adapt is the first agent in the AI Workflow Agents category to receive a Verified badge. SOC 2 Type I confirmed, pricing and integrations independently validated.

Added Adapt to AI Workflow Agents. Freemium ($0 to start), SOC 2 Type I certified, 16 native integrations. First verified listing in the AI Workflow Agents category.

Launched 8th category: AI Customer Success Agents. Added 10 agents covering the CS platform space: Gainsight, ChurnZero, Vitally, Planhat, Totango, Custify, Akita, ClientSuccess, EverAfter, and Staircase AI. Full category page, editorial listings, and guide live at /ai-customer-success-agents.

Added Tabstack to AI Workflow Agents (browser automation). Mozilla-built API for AI agents to extract structured web data and automate browser actions, with schema-as-contract reliability and 30-second MCP setup. Free tier; paid plans from $99/month.

Every active listing now has a sub-type within its primary category — 44 types across 7 categories. Rebuilt the Similar Agents algorithm to score sub-type match as the primary signal, weighted by tag rarity. Apollo.io now correctly recommends Cognism, Avina, and Orange Slice instead of CRMs. Seven listings moved to their actual competitive home (Mastra, Tines, Cognee, Mem0, Viktor, Contextpool, Claude for PowerPoint into AI Workflow Agents).

Updated 8 high-impression listings to the full editorial standard: Apollo.io, Instantly.ai, Mem0, Cognee, OpenHands, Bardeen, Cursor, and Viktor. Verified sameAs URLs added (G2, Crunchbase, Product Hunt, LinkedIn, GitHub, Wikipedia where applicable), meta titles trimmed under 60 characters, descriptions trimmed under 160. Affiliate listings (Apollo, Instantly, Lemlist) updated to the same editorial standard with "Not affiliated." disclosure preserved across all listings regardless of revenue model.

Added 2 new agents: Lightfield (AI Sales Agents -- autonomous CRM reconstruction from email and meeting history, $79/month) and Inrō (AI Marketing Agents -- Meta-verified Instagram DM automation agent, freemium from €12.99/month).

Completed full audit across 270+ active agents. Corrected 55+ listings including pricing model updates (Plandex now free, Clarm now freemium), integration corrections across 20+ agents, and deactivation of stale listings.

Updated MCP integration data across 3 agents: ContextPool now lists Windsurf, Kiro, and MCP as native integrations. Goose adds MCP support alongside VS Code and JetBrains. Clarm adds MCP alongside Salesforce, Discord, and GitHub.

Launched 7th category: AI Workflow Agents. Added 6 new agents (Axiom.ai, Browse AI, n8n, Beam AI, Zapier, Make) and moved 4 existing agents (Skyvern, Bardeen, Lindy, Relevance AI) into the new category. Full category page, guide, definition, and alternatives pages now live.

Published 2 new editorial stacks: Open Source TypeScript Agent Stack (Mastra + Cognee + Stagewise) and Small Business Social Media & Lead Gen Stack (FeedHive + Apollo.io + Lemlist).

Updated OpenAI Codex listing to reflect major 2026 upgrades: multi-agent parallel execution, desktop app for macOS and Windows, computer use, in-app browser, GPT-5.4 model, 90+ new integrations, and Codex Security launch.

Added 8 new agents: Stagewise (AI coding, YC S25), HackerEarth OnScreen (AI HR), Viktor (AI coworker), Mastra (TypeScript agent framework, YC S25), Cognee (AI memory engine), Mem0 (AI memory layer), Bardeen (browser automation), FeedHive (AI social media).

Removed deprecated and acquired agents including Thankful (acquired by Gladly 2023), Kompas AI (shut down Dec 2025), and ~30 duplicates and inactive listings.

Added 4 new agents: Ovren (AI coding), FuseAI (AI sales), Clarm (AI sales), and Tines (workflow automation). All sourced from recent Product Hunt launches.

Fixed editorial ratings not displaying on agent listings where no user reviews existed yet. All 250 agents now show correct star ratings.

Full editorial audit complete across all 250 active agents — every listing now has an editorial rating, pros, limitations, and best_for field populated.

Pricing model updated from free to freemium to reflect updated notebook limits.

Pricing updated from freemium to custom — free and pro tiers discontinued July 2025, now enterprise and custom only.

Added to the AI Agent Index across the Sales Agents category.