Every tool, one dashboard.
AI agents and orchestration. Data processing and research. Marketing and sales automation. Multi-tenant management. 25 modules built to work together — and every card below shows the plan it unlocks on.
Build & Deploy AI
Multi-agent workflows, RAG chatbots, voice agents, and full-stack app generation with one-click deployment.
AI Agents
Build multi-agent teams that collaborate on tasks. Each agent gets its own system prompt, knowledge base, and set of API tools. Five orchestration patterns — sequential, parallel, hierarchical, debate, custom — cover single-agent Q&A through multi-step research. The scheduler runs agents on a cron, and spawn chains let one agent kick off others.
- 5 orchestration patterns (sequential, parallel, hierarchical, debate, custom)
- Budget, Standard and Premium model tiers — pick the quality each agent needs
- 72 API integrations wired in as agent tools
- Task scheduler with Redis-locked cron execution
- Agent spawning chains with result collection
Assistant Builder
Create RAG chatbots that answer from your documents. Upload PDFs, DOCX, or paste text -- the system chunks, indexes, and retrieves relevant context on every query. Three modes: standard (simple chat), enhanced (web search + tools), and agentic (full multi-agent). Embed the widget on any site with a script tag.
- Standard, enhanced, and agentic conversation modes
- Knowledge base with chunking, keyword extraction, and vector search
- Embeddable widget with white-label CSS variables
- 219 agent tools your assistant can actually call
- Persistent memory with entity and fact extraction
Chat Integrations
Connect your assistants to 15 messaging platforms. Each integration handles webhook verification, message routing, and platform-specific formatting automatically. One assistant can serve multiple platforms at once. All tokens are Fernet-encrypted at rest.
- Telegram, Discord, Slack, WhatsApp, Microsoft Teams
- Matrix, Signal, Nextcloud, WebChat, Nostr
- Zalo OA, Zalo Personal, iMessage, Tion
- Per-platform webhook verification and message formatting
- Encrypted token storage with Fernet
Voice Agents
Real-time AI voice calls that pull from your knowledge base. Callers talk to an AI agent that retrieves answers from your documents and maintains conversation context with natural turn-taking. Standard and Premium voice tiers available.
- Sub-second voice responses with natural conversation
- Carrier-grade phone service for inbound and outbound calls
- Knowledge base context injection per call
- 8 voice options with natural turn-taking
- Credit-tracked per minute of call time
NeuroGen
Describe what you want built and NeuroGen assembles a team of AI agents to do it. The pipeline runs Plan, Assemble, Execute, Review, Synthesize -- generating working code across 20 framework targets. Docker sandboxing isolates execution. Selenium self-tests the output. Deploy to a wildcard subdomain with one click.
- 20 build targets: React, Vue, Express, FastAPI, Flask, Tailwind, Expo and WordPress themes
- Docker sandbox with isolated execution per session
- Selenium self-testing validates generated projects
- Wildcard subdomain deployment (*.apps.neurogen.space)
- Full-stack wiring with shared API contracts and JWT auth
Standalone Chat
2-panel chat interface you can embed or share as a link. Sidebar navigation on the left, full-width conversation area on the right — the familiar two-pane chat layout. Themeable via CSS variables for white-label deployments.
- 2-panel layout: sidebar with conversation history + main chat area
- White-label theming via CSS variables (colors, logo, branding)
- GitHub-style code blocks with syntax highlighting and copy button
- 322 TTS voices across 75 languages for spoken replies
- Mobile responsive with sidebar overlay mode
Data & Research
File conversion, web extraction, academic paper search, business analysis, and transcript processing.
File Processor
Convert documents into structured data for LLM training or RAG ingestion. Drop in PDFs, DOCX, XLSX, images, code files, or archives. The processor preserves document structure (headers, tables, lists) and outputs JSON, JSONL, Markdown, or CSV. Batch mode handles up to 1,000 files per job with live progress via SocketIO.
- 19 document formats plus 25 code file types: PDF, DOCX, XLSX, PPTX, images, archives
- OCR for scanned documents and images
- Batch processing up to 1,000 files per job
- Output as JSON, JSONL, Markdown, or CSV
- Structure preservation for headers, tables, and lists
Web Scraper
Extract content from websites with single-page, bulk, or sitemap modes. The crawler respects robots.txt, handles JavaScript-rendered pages, and follows pagination up to 5 levels deep. Output as JSON, Markdown, or CSV. Up to 500 pages per crawl job.
- Single page, bulk URL, and sitemap crawl modes
- JavaScript rendering for dynamic sites
- Up to 500 pages per crawl, 5 levels deep
- robots.txt compliance built in
- Output in JSON, Markdown, or CSV
Academic Search
Search across 17 academic databases simultaneously -- CrossRef, Semantic Scholar, arXiv, PubMed, CORE, and more. Results come back with relevance scoring, and the system can auto-download available PDFs. Generate citations in APA, MLA, Chicago, Harvard, or BibTeX format directly from the results.
- 20 academic sources searched in parallel
- Relevance scoring across all result sources
- Automatic PDF download when available
- Citation generation in APA, MLA, Chicago, Harvard, BibTeX
- Citation network analysis for related papers
Business Intelligence
Upload financial statements, market data, or competitive research and get structured analysis back. Three specialized AI agents handle financial modeling, market assessment, and competitor analysis. Outputs include revenue trends, SWOT breakdowns, and entity relationship extraction.
- 3 specialized AI agents (financial, market, competitive)
- Revenue trends, cost structure, and profitability analysis
- SWOT analysis and competitor mapping
- Entity extraction for companies, people, and relationships
- Live progress updates during processing
Playlist Downloader
Pull transcripts and metadata from YouTube playlists in bulk. Useful for building training datasets, populating knowledge bases, or archiving video content as text. Each video's transcript, title, description, and timestamps export as JSONL for direct use in AI pipelines.
- Batch transcript extraction from entire playlists
- Metadata capture: titles, descriptions, timestamps, tags
- JSONL export formatted for LLM training and RAG
- Handles hundreds of videos per job
Marketing & Sales
Social media management, ad campaigns, communications, and multi-step sales funnels with course delivery.
Social Media
Manage 11 platforms from one dashboard: Facebook, Instagram, Threads, YouTube, LinkedIn, Pinterest, X/Twitter, TikTok, Reddit, Bluesky and Google Business Profile. The AI content engine generates captions, hashtags, images and video. Schedule posts, run campaigns with UTM tracking, set up lead capture rules, and view cross-platform analytics.
- 11 platforms with OAuth2 token management and auto-refresh
- AI image and video generation, on the Standard or Premium model tier
- Unified inbox for comments and DMs with AI reply suggestions
- Lead capture rules with keyword triggers and engagement thresholds
- Campaign management with UTM tracking and content calendar
Ads Manager
Run paid campaigns across 6 ad platforms from a single interface: Meta, Google, LinkedIn, Pinterest, Reddit, and TikTok. The AI creative engine generates ad copy and visuals. Audience management handles targeting across platforms, and the analytics dashboard shows spend, conversions, and ROAS side by side.
- 7 ad channels: Facebook, Instagram, Google, LinkedIn, Pinterest, Reddit, TikTok
- AI-generated ad creative (copy and visuals)
- Cross-platform audience management and targeting
- Unified analytics with spend, conversions, and ROAS
- Campaign creation, budgeting, and scheduling from one screen
Communications
Browser-based phone system with dialer, SMS inbox, email, and autoresponder sequences. Calls and texts go out over carrier-grade phone service with STIR/SHAKEN compliance. The autoresponder handles multi-channel drip campaigns with conditional branching based on opens, clicks, and non-responses. Lead scoring weights activity across all channels.
- Browser dialer with call recording, transfer, and scoring
- SMS inbox with templates and conversation threading
- Authenticated email delivery with autoresponder sequences
- Conditional sequence branching (open/click/no-response)
- Lead scoring across calls, SMS, email, and social activity
Sales Funnels & Courses
Build multi-step sales funnels with 8 funnel types and 10 step types. Products can be digital, physical (with 5 shipping carriers), chatbot access tokens, or online courses. Stripe Connect handles payments. A/B test any step, enable one-click upsells with saved payment methods, and map custom domains with auto-SSL via Let's Encrypt.
- 8 funnel types, 10 step types, per-step conversion analytics
- A/B testing on any funnel step with traffic splitting
- One-click upsells using saved Stripe payment methods
- Course builder with drip-gated modules, video player, progress tracking
- Custom domains with DNS verification and Let's Encrypt SSL
Enterprise & Intelligence
Multi-tenant organizations, knowledge retrieval, and specialized document analysis.
Organizations & RBAC
Create organizations with role-based access control. Five roles (owner, admin, manager, member, viewer) map to 12 permissions that control who can do what. Organizations get their own credit pools that delegate to members and customer accounts. White-label branding lets you set custom colors, logos, favicons, and your own domain.
- 5 roles x 12 permissions with org-scoped data isolation
- Credit delegation: org pools, member allocations, customer accounts
- Customer portal with separate login and credit tracking
- White-label branding (colors, logo, favicon, custom CSS, domain)
- Analytics dashboard with usage trends, model breakdown, and member activity
NeuroGen Cortex
Solves context rot — the silent accuracy killer in every other AI platform. Backed by peer-reviewed long-context research, NeuroGen Cortex maintains AI accuracy across unlimited document sizes. Structural reading, adaptive retrieval for complex questions, and persistent memory that learns across sessions.
- Process documents 100x beyond standard context limits
- Adaptive retrieval finds answers other AI misses
- Complex questions automatically decomposed and answered
- Persistent memory across sessions — your AI learns over time
- Confidence-gated answer verification before delivery
Legal Discovery
Upload legal documents for automated review. The system runs multiple analysis passes: privilege detection identifies attorney-client communications, contract analysis extracts terms and deadlines, risk assessment flags compliance issues, and hot document identification surfaces fraud and liability indicators. Export results as structured PDF reports.
- Attorney-client privilege detection
- Contract analysis: terms, obligations, deadlines, parties
- Risk assessment for litigation and compliance exposure
- Hot document flagging for fraud, liability, and misconduct
- PDF export with structured findings and citations
New since August
Work that landed after this page was last written. Described as what it can do — not as what anyone has proven with it yet.
Communications MCP Server
Your assistants can read the communications module through a governed MCP server rather than a raw toolkit — with a tier floor, a rate limit and a response-size cap enforced before any tool runs, and per-agent allow-listing of which servers an agent may reach at all.
- Read contacts, campaigns, enrolments and call outcomes from an assistant
- Tenant identity is injected by the executor, never taken from tool arguments
- Adds no new outbound sending, deliberately — one enrolment tool, and it sends nothing itself
In-Call Actions
The voice agent can do things during the call, not just talk about them: check what is free, book the appointment, hand the call to a person, look up the caller's own order, or save a note to their contact. Each action is defined once and shared by the voice runtime and your agents, so they cannot drift apart.
- Books against your existing appointments and your connected calendar's busy times
- Transfer target is set by you in configuration — the model never picks a number
- Order lookup is always the caller's own; it cannot be pointed at anyone else
Follow-Up Branching
An email follow-up sequence can now take a different path depending on what the contact did — opened, clicked, or neither — with the branches edited in the campaign UI instead of in a config file. Steps can also wait for a condition before they run.
- Branch on engagement; jump forward to a later step in the sequence
- Authored in the campaign editor, no JSON to hand-write
- One enrolment per contact per campaign, enforced in the database
WordPress Themes
WordPress is now one of the 20 build targets. Describe the site and the builder generates a complete theme — templates, stylesheet and functions — which you download and install on your own WordPress host.
- Generated as a theme you own, not a hosted page on our platform
- Same prompt-to-preview loop as the other build targets
Machine-Readable Pages
Generated sites ship with the structured data an answer engine needs to quote them: JSON-LD for the organisation, the FAQ and the article, plus robots.txt and an llms.txt describing the site to crawlers. Each page is scored on what is actually present — a resolving canonical, a parseable FAQ, a real postal address.
- JSON-LD, robots.txt and llms.txt emitted with the build
- Scored on what the page contains, and on NAP consistency across it
- We score presence, not outcomes — no ranking or citation is promised
Start building on the free plan.
6 modules and 100 credits, no card. The app builder and one-click deploy open on Starter at $47; the AI voice agent and funnels on Professional.