Oct 2026
Multi-domain serving
One deployment answers on several hostnames, with the self-reference guard covering all of them.
Agentify your product.
An assistant that works inside your product, not next to it. It runs under your name, in your branding, acting with each user’s own permissions.
Three things the assistant does, in order: it highlights the real control in your product rather than describing where to find it; it calls your tools as the signed-in user, with that person’s own permissions; and anything that moves money waits for an explicit approval before it runs.
A demonstration: a user opens the assistant inside their application from an Ask AI button, which prefills the question “Tell me about this screen”. The assistant describes the page, then highlights the real Invite button when asked where to add a teammate. The user reports an incorrect invoice; the assistant reads it, offers the possible explanations rather than guessing, cites the billing policy it answered from, waits for approval before issuing a refund, and offers to check seat usage every month.
yourapp.com
POINTS AT THE REAL CONTROL
Not “go to Settings, then Team”. It rings the actual button in your app.
ASKS INSTEAD OF GUESSING
When a question has more than one honest reading, it offers the options rather than picking one.
ANSWERS FROM YOUR DOCS
It cites the passage it read, so the answer is checkable rather than taken on trust.
WRITES PAUSE FOR APPROVAL
Anything that moves money waits for an explicit yes — then runs as Sara, with her permissions.
The part of your product that answers back. It reads your docs, calls your APIs, and does the thing — inside the app your users are already in.
01CONNECT YOUR STACK
Your tools, already its tools. Point it at an API or an MCP server and the whole toolset imports itself — no glue code, nothing to keep in sync by hand.
The tools page of the console: a searchable list of webhook and MCP tools with call counts, one of them switched off.
Webhooks, guarded
Every outbound call is SSRF-checked and carries a stable idempotency key, so an ambiguous failure never fires a write twice.
MCP, natively
Connect any MCP server and its tools arrive disabled, so you choose what the agent is allowed to touch.
02TEACH IT YOUR PRODUCT
It answers from your words. Upload the docs, the help centre, the policies. Hybrid search and reranking pick the passage, and every answer can show which one.
The knowledge page: ingested sources on the left, and on the right the passage retrieved for one answer with the matched phrase marked.
Per-user knowledge
Each user can carry private documents only they can retrieve, scoped by signed identity rather than by a guessable id.
Skips when empty
No knowledge base, no retrieval step — the agent stops paying for a search that was never going to return anything.
03PUT IT IN YOUR APP
One script tag. It renders inside your page rather than in an iframe, in your branding, streaming answers token by token.
An embed snippet beside the page it produces: the widget docked in a customer app, with the Export button ringed.
Points at your UI
Because it lives in your page, it can ring the actual control instead of describing where to find it.
Voice in and out
Dictate instead of typing and have answers read back. Browser-native, so it costs nothing and works with every model.
04LET IT ACT FOR EACH USER
As them, not as everyone. Signed per-user identity and an encrypted credential vault, so it does things against your systems with that person’s own permissions.
A signed user_context payload with line numbers, and the three things the signature enforces.
Writes pause
Anything that moves money or data waits for an explicit confirmation before it fires.
Budgets and limits
Per-project rate limits and a monthly ceiling, enforced server-side rather than hoped for.
05SEE EVERYTHING IT DOES
No black box. Every turn, tool call, retrieval and token is logged and priced, so you can answer where the bill came from.
The analytics page: turns, tool calls, cache reads and spend, a seven-day chart with one column’s detail open, and the last three tool calls.
Cost, broken down
Input split across tool schemas, system prompt, history and the message itself — the parts add up to the total exactly.
Replay any turn
Open a single conversation and see the exact context the model was given, not a reconstruction of it.
The platform underneath
One agent runtimefor every product you ship.
It logs itself
Every turn, tool call and token, priced and searchable.
Your tools talk
Webhooks and MCP servers arrive as one toolset.
It knows the user
Signed identity, their permissions, their credentials.
It asks first
Writes pause for approval before anything moves.
It stays in budget
Rate limits and a monthly ceiling, enforced server-side.
AGENTIFYS MCP
Set it up from your editor.Not from a settings page.
Point your coding agent at one URL and it provisions the whole project — prompt, user schema, tools, policies, connected MCP servers, limits and scheduling. Twelve tools, a scoped token, and no secret is ever an argument.
$ claude mcp add --transport http agentifys https://agentifys.ai/mcpA terminal transcript: a coding agent creates the project, configures signed identity, sets the user schema and prompt, adds a tool, connects an MCP server whose nine tools import disabled, verifies the project, and finishes with one manual step — pasting the model key, which no tool accepts as an argument.
REST. Widget. MCP.Build on it however you work.
A documented HTTP API, an embeddable widget with a JavaScript handle, and the MCP server above. Three ways in, and underneath all three the same runtime, the same tools, and the same per-user credentials — pick whichever fits how you build.
Measured, not estimated
Production-grade from the first message. These are numbers off our own traffic, not a benchmark.
93%
of prompt tokens served from cache on live projects
80%
lower cost per model call once caching is on
1
script tag to embed the whole agent
0
shared keys — every project spends its own
Better every week. Small, frequent, and written down.
View allSep 2026
Prompt caching
Repeated tool schemas serve from cache — about 80% off a typical model call.
Sep 2026
Point at a control
Configure screen elements in a setup screen instead of hand-editing the system prompt.
Sep 2026
Per-component cost
See which part of the context your spend came from: schemas, prompt, history or the message.