# agent4.io > The commercial platform for pre-sales-consulting and customer-follow-up AI agents. Businesses run an agent that answers from their own knowledge (not the open internet), qualifies and follows up with their customers across web, WhatsApp and Telegram, and — for games and creator products — gives a character a memory that persists per end-user. Multi-tenant, with per-user encrypted memory, MCP tool-calling, skills, knowledge bases, scheduled follow-ups, and output-side moderation. ## Key pages - [Pricing](https://agent4.io/pricing): plans and token quotas - [Why agent4.io](https://agent4.io/why): what makes it different - [Docs](https://agent4.io/docs): product documentation ## Industry & use-case solutions - [Litigation and family law](https://agent4.io/solutions/legal): An agent that meets people at the worst moment of their year, takes the intake your firm actually needs — parties, dates, jurisdiction, urgency — books the consultation, and chases the documents, grounded in your firm's own practice notes rather than the open internet. - [Company formation and incorporation agents](https://agent4.io/solutions/company-formation): An agent that answers jurisdiction, fee and document questions from your own price sheets — not the open internet — qualifies cross-border founders, runs KYC collection to a complete file, and reminds every client before every annual deadline, across your site, WhatsApp and Telegram. - [K12 schools and tutoring](https://agent4.io/solutions/k12): Build subject teachers on top of your own curriculum, give every student a private space where their weak points accumulate, and let the platform chase the revision. Not a chatbot per school — a teacher per subject, per student, with the memory that makes teaching different from answering. - [Immigration](https://agent4.io/solutions/immigration): An agent that screens eligibility against the routes you actually file, explains the process in the applicant's own language, and holds the document chain together across the months — or years — a case takes to reach a decision. - [Insurance](https://agent4.io/solutions/insurance): An agent that explains coverage in plain language, pre-qualifies quote requests, answers policy questions from your own product sheets, and never lets a renewal date slip past a customer. - [Lending](https://agent4.io/solutions/lending): An agent that answers rate and eligibility questions from your own program sheets, pre-qualifies borrowers, walks them through document checklists, and follows up on repayments and renewals on schedule. - [Study abroad](https://agent4.io/solutions/study-abroad): An agent that answers program and visa questions from your counseling playbooks, matches students to realistic options, and keeps every application timeline — deadlines, documents, interviews — moving on schedule. - [Software and IT services](https://agent4.io/solutions/software): An agent that runs the discovery you currently spend senior engineering hours on — drawing out what the prospect actually needs, checking it against what you actually build, and handing your team a scoped brief instead of "they want an app". - [Hospitality and personal services](https://agent4.io/solutions/hospitality): An agent that remembers how each guest likes to be looked after — the usual drink, the corner table, the therapist they ask for, the dates that matter — books them in, and briefs your team before they walk through the door. - [Apparel and fashion retail](https://agent4.io/solutions/apparel): An agent that remembers what size someone actually took last time, answers fabric and washing questions from your own care guides rather than the open internet, and follows up when a favourite is back in stock — across your site, Telegram and WhatsApp. - [Game companions and narrative characters](https://agent4.io/solutions/game-companions): Give one character a private, persistent relationship with every player. It recalls what they did ten hours ago, stays inside your canon, acts through your game's own functions, and can't be jailbroken into a headline — running behind your game over an API, invisible to your players. - [Character platforms and creator tools](https://agent4.io/solutions/character-platforms): Building a product where creators bring characters to life? Build the marketplace, the payouts and the audience yourself — that's your product. The hard part is a character that actually remembers each fan, never breaks a creator's canon, and can't be jailbroken at scale. That part is infrastructure, and it's what we run. - [Online course businesses](https://agent4.io/solutions/online-courses): An agent per course that answers "is this right for me" from your own curriculum and sales pages, on the platforms your buyers already use — no website to build. It qualifies the buyer, replies in their language, remembers them between launches, and hands your team the questions your sales page never answered. - [Loan and credit application agents](https://agent4.io/solutions/loan-origination): An agent that runs the whole loan enquiry as a guided journey — collecting documents step by step, remembering where each applicant is across days and channels, calling out to your credit and KYC checks, and handing your team only the applications that are ready — while never quoting a rate or promising an outcome it isn't allowed to. - [SaaS onboarding and activation agents](https://agent4.io/solutions/saas-onboarding): An agent that runs activation as a guided journey — greeting each new signup, walking them step by step to their first real outcome, reacting to what they actually do in your product rather than what they say, and reaching them across the app, email and chat until they're set up. When they're activated, it hands them on; when they stall, it steps in. ## Agent templates (ready-made agents you can create in one step) - [Insurance quote and renewal agent](https://agent4.io/templates/insurance-quotes): A pre-sale agent for brokers and agencies that explains cover in plain language from your own policy documents, qualifies a quote request into something an underwriter can price, and reminds every client before their renewal date. - [Legal client intake agent](https://agent4.io/templates/legal-intake): An intake agent for law firms that answers process and fee questions from your own materials, screens an enquiry against the matters you take on, collects the facts a fee-earner needs before the first call, and never gives legal advice. - [Company formation agent](https://agent4.io/templates/company-formation): An agent for incorporation agencies that answers which-jurisdiction and how-much questions from your own fee schedules, qualifies cross-border founders across time zones and languages, and works through your KYC checklist until the file is complete. - [Course pre-sale agent](https://agent4.io/templates/course-presale): An agent for course businesses that answers prerequisites, time commitment, access and refund questions from your own course materials, tells the wrong-fit buyer so before they pay, and hands you a qualified lead instead of an inbox. - [Immigration intake and case-chasing agent](https://agent4.io/templates/immigration-intake): An intake agent for immigration practices that explains your process and fees from your own guidance notes, screens an enquiry against the routes you file, collects a structured profile for a caseworker, and chases the outstanding documents on a live case — without ever predicting an outcome. - [Loan pre-qualification agent](https://agent4.io/templates/loan-prequalification): A pre-sale agent for lenders and brokers that answers program and eligibility questions from your own rate sheets and guidelines, collects what a loan officer needs to assess a borrower, chases the document checklist until the file is complete, and never promises approval. - [Study abroad advising agent](https://agent4.io/templates/study-abroad-advising): An agent for study-abroad consultancies that answers programme, cost and application-process questions from your own counselling documents, collects the profile a counsellor needs before the first call, and reminds each student before every deadline — without ever predicting whether they will get in. - [Tutoring enrolment agent](https://agent4.io/templates/tutoring-enrolment): An enrolment agent for tutoring centres that answers parents' questions about levels, timetables, class sizes and fees from your own course material, books the trial lesson or teacher call, and hands anything about a specific child's difficulties to a teacher rather than assessing the child itself. - [Booking and guest-preference agent](https://agent4.io/templates/booking-concierge): A front-of-house agent for restaurants, bars, spas and treatment rooms that answers questions about your venue from your own material, collects a booking request with everything your team needs to confirm it, and remembers the preferences and occasions that make a regular feel known. - [Fit and garment-care agent](https://agent4.io/templates/fit-and-care): An agent for apparel brands that answers "what size am I in your brand" from your size charts and fit notes, answers washing and fabric questions only from your own care documentation, remembers what each customer kept and returned, and follows up when their size is back. - [Project scoping agent](https://agent4.io/templates/project-scoping): A pre-sale agent for software houses, dev agencies and IT services firms that answers capability and integration questions from your own catalogue, works a vague enquiry through the questions your engineers always ask, and hands your team a structured requirements brief — without quoting a price or committing to a date. - [Character companion agent](https://agent4.io/templates/character-companion): An agent for novelists, game studios and character platforms that lets a reader talk to one of your characters — with the persona, voice and canon you supply, a memory of each reader that persists between sessions, and the boundaries a character must never cross however the conversation goes. ## Case studies (illustrative worked examples) - [How a bridge-lending team would put intake and follow-up on autopilot](https://agent4.io/cases/2026-06-bridge-lending-pilot): A worked example of a private lending firm running an agent4.io agent as the first touchpoint on its website and WhatsApp line — answering rate questions from the firm's own program sheets, pre-qualifying borrowers, and chasing document checklists so officers start from a complete file. - [How a company-formation agency would run intake, KYC and renewals on autopilot](https://agent4.io/cases/2026-07-company-formation-agency): A worked example of a small cross-border incorporation agency running an agent4.io agent as the first touchpoint on its website and WhatsApp — answering jurisdiction and fee questions from its own price sheets, qualifying overseas founders in their own language, collecting KYC to a complete file, and reminding every client before every annual deadline. - [How a K12 provider would give every student a teacher per subject](https://agent4.io/cases/2026-07-k12-tutoring-platform): A worked example of an education company building its own teaching product on agent4.io — a virtual teacher per subject per school, each student's weak points accumulating in a space only they can reach, and revision that arrives on a schedule instead of waiting for the student to open the app. - [How a restaurant would answer every table over a Telegram QR code](https://agent4.io/cases/2026-07-restaurant-telegram-orders): A worked example of a busy restaurant putting an agent4.io agent behind a Telegram QR code on every table — answering menu, allergen, and wait-time questions in the guest's own language, remembering regulars by their Telegram account, and passing anything it can't answer to a staff member with the table number attached. - [How a study-abroad agency would stop losing leads between time zones](https://agent4.io/cases/2026-07-study-abroad-agency): A worked example of a consultancy serving students across three time zones running an agent4.io agent on its website and Telegram — answering program and visa questions from the agency's own counseling notes, profiling students, and running per-school deadline reminders. - [How a SaaS team would turn onboarding into a journey that finishes itself](https://agent4.io/cases/saas-activation-journey): A worked example of a B2B SaaS running an agent4.io storyline as its activation flow — greeting each signup, walking them to their first real outcome, advancing on the product's own events rather than what users claim, and stepping in when an account stalls, across the in-app assistant and email. ## Concepts (glossary) - [Agent harness: the layer that turns a model into a dependable agent](https://agent4.io/concepts/agent-harness): The engineering layer wrapped around a language model — retrieval, tools, skills, memory and guardrails — that turns a general model into a specific, dependable agent. The model supplies fluency; the harness supplies grounding, boundaries and the ability to act, and stays the same when the model is swapped. - [Agent-to-Agent (A2A): when your customers' AI assistants do business with your agent](https://agent4.io/concepts/agent-to-agent): A machine-to-machine channel where a customer's personal AI assistant talks directly to a business's agent — consulting it and submitting a structured request — instead of a person typing into a chat box. agent4.io handles the pre-processing half and hands back a summary, while a human confirms anything irreversible. - [AI agent vs LLM: what is the difference?](https://agent4.io/concepts/agent-vs-llm): The model is the engine that turns text into text; the agent is everything built around it — retrieval, tools, memory, boundaries and channels. Models are replaceable parts inside an agent. - [What are ask forms (forms inside the chat)?](https://agent4.io/concepts/ask-forms): A small form the agent composes mid-conversation — choices, a number, a date — rendered as something to tap rather than a paragraph of questions to answer by typing. - [How to build a knowledge base for an AI agent](https://agent4.io/concepts/building-a-knowledge-base): Most of what an agent needs to know is already written down — on your website, and in the PDFs your team already sends customers. Importing from a URL covers the public half; uploading covers the rest. - [What is a business AI agent?](https://agent4.io/concepts/business-agent): An agent that works for one specific business — answering only from that company's own documents, staying inside a scope its owner defines, remembering each customer across visits, and accountable for what it says. Not a general chatbot, but a member of that business's team. - [Fine-tuning vs RAG: when you actually need to fine-tune](https://agent4.io/concepts/fine-tuning): Continuing a model's training on examples of the behaviour you want, so that behaviour becomes its default instead of something you ask for on every request. It changes how a model works, not what it knows. - [agent4.io as headless software](https://agent4.io/concepts/headless): Headless software exposes everything over an API, MCP and skills so that an AI agent — not only a human clicking a UI — can operate it. You describe the outcome; your agent, handed exactly the right skills, drives the features precisely. - [Knowledge base, tools or Skill — which does what](https://agent4.io/concepts/how-the-pieces-fit): Three different jobs that get confused for one — the knowledge base holds facts, an MCP tool performs a calculation you cannot let a model improvise, and a Skill is the procedure that decides which to reach for. - [What is a knowledge base for an AI agent?](https://agent4.io/concepts/knowledge-base): The documents an agent may answer from. agent4.io keeps two kinds — your company's vetted material, and each customer's own files — retrieved together but never confused for one another. - [What is MCP (Model Context Protocol)?](https://agent4.io/concepts/mcp): The Model Context Protocol — an open standard for connecting an agent to external tools and data, so a system exposes its capabilities once instead of once per AI product. - [What is a page playbook (page-aware AI chat)?](https://agent4.io/concepts/playbooks): A per-page briefing for the agent — background it should already know, an opening line, and a few suggested questions — chosen automatically from the URL the visitor opened the chat on. - [What are scheduled tasks for an AI agent?](https://agent4.io/concepts/scheduled-work): Work an agent commits to for a specific future moment — a follow-up it decides to send, or a recurring reminder — executed on time by a scheduler rather than waiting for someone to speak first. - [What is an agent Skill?](https://agent4.io/concepts/skills): A named capability pack — a short description of when to use it, fuller instructions, and the tools it unlocks — that the agent loads only when it becomes relevant. - [What is an agent's soul (system prompt)?](https://agent4.io/concepts/soul): The standing description of who an agent is — its character, tone and non-negotiables — held separately from the job it happens to be doing. - [What is a space (agent workspace)?](https://agent4.io/concepts/spaces): The isolated container holding one customer's conversations, documents and memory — enforced at the database level, so one customer's material cannot surface in another's conversation. - [Storylines: stateful, graph-orchestrated agents that deliver a process](https://agent4.io/concepts/storylines): A storyline is a directed graph an agent follows with each end user — every node is a step (its own task, knowledge and tools), exits carry conditions, and each person's progress, profile and notes are saved and resumed across sessions and channels. It turns an agent from an assistant that does point-tasks into one that can deliver a multi-step service itself. - [What is an AI agent?](https://agent4.io/concepts/what-is-an-agent): Software that holds a conversation, looks things up in your documents, and takes actions on your systems — deciding for itself which to do next and following a task through, step by step, rather than only replying to one message at a time. ## Recent changes - [2026-07-26 — Storylines: agents that run a whole process, not just one task](https://agent4.io/changelog): Build a graph an agent follows with each person — steps with their own knowledge and tools, branching conditions, a profile that accumulates, and progress saved and resumed across sessions and channels. - [2026-07-25 — Human-readable links for your agents](https://agent4.io/changelog): Give any agent a short, memorable public link — chat.agent4.io/t/your-workspace/menu — instead of an opaque share token, without changing the agent's name. - [2026-07-25 — "Improve it" is now a guided wizard](https://agent4.io/changelog): The dashboard's "Improve it" no longer starts from a blank box. You pick what to improve, the AI reads it and offers concrete suggestions to tick from, and you review the exact before/after before anything is written. - [2026-07-25 — A writing assistant for your agent's prompts](https://agent4.io/changelog): Next to every big text field in the console — an agent's persona and task, a skill's description and instructions, a knowledge base's notes — there's now an "Improve with AI" button that rewrites what you've written into a sharper instruction for the AI. - [2026-07-24 — Sign up with X](https://agent4.io/changelog): You can now create an account with X (Twitter) in one click, alongside Google and GitHub — with a short email step that keeps your account reachable and secure. - [2026-07-24 — Buy extra tokens on any plan](https://agent4.io/changelog): Top up your token balance without a subscription — from 1M tokens, card entered at checkout — so the free plan is no longer a hard ceiling. - [2026-07-23 — A command-palette "Ask" bar — a second way to embed your agent](https://agent4.io/changelog): Alongside the corner bubble, you can now drop a slim "Ask a question" bar that opens a centered command palette on ⌘K — the same one line of script, the same agent behind it. - [2026-07-23 — Customers can now upload documents — and they land in your inbox](https://agent4.io/changelog): A visitor can attach a photo or a PDF in the chat — a diploma, a bank statement, an ID — and it appears on the inbox record for your team to review. - [2026-07-23 — Your dashboard now shows end-user usage — and a 24-hour token view](https://agent4.io/changelog): The overview tells you how many of your plan's end users you've used, and the token chart now switches between the last 24 hours, 7 days, and 30 days. - [2026-07-23 — Your agent can recommend products — with photos](https://agent4.io/changelog): A new product-catalog knowledge base lets your agent suggest dishes, clothing or any item and show the photo right there in the chat. - [2026-07-22 — See what each agent is still missing](https://agent4.io/changelog): Every agent now carries a checklist that is answered from real data — including the two failures that were previously completely silent. - [2026-07-22 — Change one setting without resending the rest](https://agent4.io/changelog): The API can now update a single field on an agent or skill, and every change is kept in a history you can roll back to. ## Documentation - [Authentication](https://agent4.io/docs/api/authentication): The header schemes for tenant, end-user, and proxy-mode API access. - [Chat API](https://agent4.io/docs/api/chat-api): Drive conversations programmatically — synchronous or streaming, with retrieval sources in every reply. - [Set it up with your coding agent](https://agent4.io/docs/getting-started/agent-native): You don't have to read all of these docs. Hand them to your AI coding agent — via MCP and Skills — and it will integrate, configure, and build on agent4.io for you. - [Platform overview](https://agent4.io/docs/getting-started/overview): The concepts behind agent4.io — tenants, agents, spaces, knowledge, and channels — and how they fit together. - [Quickstart](https://agent4.io/docs/getting-started/quickstart): From an empty tenant to a working agent on your website in about fifteen minutes. - [Where things are](https://agent4.io/docs/getting-started/where-things-are): The new-here map — what an agent, a skill and a knowledge base each are, and where in your dashboard you change what your agent says, add your own data, and turn features on. - [One agent, every channel](https://agent4.io/docs/how-it-works/channels): How the web widget, Telegram and WhatsApp reach the same agent with the same memory — identity resolution, the shared control flow, and what stays channel-specific. - [Scheduled follow-ups](https://agent4.io/docs/how-it-works/follow-ups): How an agent schedules work for later and delivers the result back to the channel it came from — durable tasks, atomic claiming, and recovery. - [Per-customer memory](https://agent4.io/docs/how-it-works/memory): How conversations become durable facts about a customer — derivation, deduplication, importance, and what gets recalled. - [Running on open models](https://agent4.io/docs/how-it-works/open-model-compatibility): How the model layer stays separate from the agent layer — capability probing, multi-backend routing, and the four ways small open models break on tool calls that the gateway handles for you. - [Grounded answers](https://agent4.io/docs/how-it-works/retrieval): How a document becomes something an agent can answer from — chunking, the context header, query rewriting, and the relevance floor. - [Private by design](https://agent4.io/docs/how-it-works/security): Isolation enforced in the database, field-level encryption with per-user keys, and an honest account of what the end-to-end layer does and does not give you. - [Tools and MCP](https://agent4.io/docs/how-it-works/tools): How an agent goes from answering to acting — the tool loop, context budgets, and connecting your own systems over MCP. - [Agent Interface (Agent-to-Agent)](https://agent4.io/docs/integration/agent-interface): Expose your business agent to other agents. A customer's AI assistant can discover it, consult it, and submit a structured, pre-filled request over MCP and A2A — while a human confirms anything irreversible. - [Embed the web widget](https://agent4.io/docs/integration/embed-widget): Put your agent on any website with one script tag — bubble, window, and full-screen modes included. - [Ready-made end-user UI](https://agent4.io/docs/integration/end-user-ui): Not a developer? Hand your customers a complete, hosted chat experience — no code, no embedding. A full end-user interface you can share as a link. - [Page playbooks](https://agent4.io/docs/integration/page-playbooks): Tell the agent which page the visitor is on, so it opens the conversation already knowing what they came for. - [Records — where what the agent collects ends up](https://agent4.io/docs/integration/records): Your agents file enquiries, document checklists, booking requests and escalations into a queue you work through, each one linked back to the conversation that produced it. Over MCP, or over plain HTTP if you have your own stack. - [Telegram & WhatsApp](https://agent4.io/docs/integration/telegram-whatsapp): Give each agent a native bot presence in the messengers your customers already use. - [White-label & custom apps](https://agent4.io/docs/integration/white-label): Go beyond the widget. Drive agents from your own product over the API to build fully custom, white-labeled experiences — support desks, in-app copilots, even game characters. - [Put your agent on Telegram](https://agent4.io/docs/tenant-guide/connect-telegram): A plain, step-by-step guide to giving your agent a Telegram bot — no coding, about five minutes. - [Testing and changing an agent](https://agent4.io/docs/tenant-guide/keeping-an-agent-honest): Run a made-up visitor at your agent to find out whether its limits hold, then change what it does by describing the change and ticking the parts you agree with. - [Workspace ID, API key, tool token & console](https://agent4.io/docs/tenant-guide/keys-and-tokens): What each credential on the Connections screen is for — your workspace ID, the API key, the tool token behind the record tools, and console access — and which ones you actually need. - [Quotas & usage](https://agent4.io/docs/tenant-guide/quotas-and-usage): What the monthly token quota covers, what happens at the limit, and where to watch consumption. - [Spaces & knowledge](https://agent4.io/docs/tenant-guide/spaces-and-knowledge): How isolation works, and how to organize documents so retrieval stays sharp. - [Build a storyline](https://agent4.io/docs/tenant-guide/storylines): How to turn an agent into a multi-step, stateful process — nodes, exit conditions, dimensions, third-party callbacks, versioning and migrating in-flight users. - [Token usage — what counts](https://agent4.io/docs/tenant-guide/token-usage): What the "pre-processing" and "generation" bars on your usage chart actually mean, what consumes tokens and what doesn't, and how to read usage over time and by agent. ## Cookbook (recipes for building on agent4.io over MCP — which tool to call) - [Build and populate a knowledge base](https://agent4.io/cookbook/build-a-knowledge-base): Create a KB, add text and files, and verify retrieval before mounting it — over MCP. - [Compile a flow into a Storyline](https://agent4.io/cookbook/build-a-storyline): Turn a multi-step process into a stateful directed graph — create, validate, publish. - [Create a grounded support agent](https://agent4.io/cookbook/create-support-agent): Create an agent with a soul, a task with boundaries, tools and a knowledge base — over MCP. - [Design principles — read this first](https://agent4.io/cookbook/design-principles): How to configure agents that actually work — one job per agent, few skills, grounded, verified. - [Get started with the agent4.io agent skills](https://agent4.io/cookbook/getting-started): Connect your coding agent to agent4.io over MCP, install the skills, and run your first recipe. - [Author a load-on-demand skill](https://agent4.io/cookbook/create-a-skill): Package a capability the model pulls only when relevant — create it and attach it to an agent. - [Modelling a process as a Storyline — the right way](https://agent4.io/cookbook/modelling-a-process-as-a-storyline): The paradigm and pitfalls for turning a real business process into a Storyline that behaves. - [Let your agent query your knowledge base](https://agent4.io/cookbook/query-your-knowledge-base): Search a tenant knowledge base over MCP — authenticated, private by default. - [Configure a page playbook (page-aware chat opener)](https://agent4.io/cookbook/configure-page-playbooks): Give the agent a per-page briefing — background, an opening line and suggested questions — chosen automatically from the visitor's URL, over MCP. - [Parallel steps — do several things, then continue](https://agent4.io/cookbook/parallel-steps): An AND-join: several sub-flows run in any order, each with its own checks; continue only when all are done. - [Deleting things — point the user to the console](https://agent4.io/cookbook/deleting-things): Deletions aren't MCP tools; tell the user the exact console screen and button instead. - [Connect your own MCP server](https://agent4.io/cookbook/connect-your-mcp-server): Register a remote MCP server so your agents can call its tools — via the console or REST. - [Built-in tools (and the system MCP)](https://agent4.io/cookbook/built-in-tools): The tools the platform ships — lead capture, documents, follow-ups, web search — plus the auto-installed time/weather/location server. - [Check usage and quota](https://agent4.io/cookbook/check-usage-and-quota): Read token usage and the monthly quota over MCP — by agent, user, space, day or model — to answer "am I close to my limit?" and "what's driving the cost?". - [Look up users and their sessions](https://agent4.io/cookbook/inspect-users-and-sessions): List your end users and their chat sessions over MCP — directory and metadata only; transcripts stay in the console, reached by a link the tools hand back. - [Brand the chat — colours and logo](https://agent4.io/cookbook/configure-branding): Match a customer's brand on a share link or embedded widget: one colour, a logo, and CSS only when the brand guide demands it.