Agent, Mate, and channel chat
Users can start a new agent session, continue lightweight work with Mate, or keep shared context inside project channels instead of copying answers between tools.
Everything an agent needs to execute work.
Eight apps, 60+ built-in tools, broad model routing, and the channels your users already use — under one account and one memory.
8
apps
60+
tools
75+
models
3
ways to run
Plan a 3-city launch trip next week and add it to my calendar.
web.search · flights
calendar.create_event ×6
Booked SF → NYC → Austin and added 6 events to your calendar.
Launch week
Mon · San Francisco — kickoff
Wed · New York — press
Fri · Austin — partners
Message the agent…
isagent.io/app
The app is a workspace, not a single prompt window.
The product workspace supports agent chat, Mate chat, channel chat, Chat/Browse/Design modes, multiple widgets, tasks, calendar, notifications, advanced isa_agent_sdk behaviors, and process orchestration in one shell.
Chat / Browse / Design modes
A session can move from normal chat to live browser research or Design Studio artifact work while preserving memory, files, and process state.
Workspace widgets
DreamForge, HuntAI, Omni, Knowledge, Data Scientist, Trading, and Maestro Workflows are working panels attached to the same agent context.
Tasks, calendar, and notifications
Conclusions can become scheduled work, reminders, calendar events, and notifications inside the app shell rather than disappearing after the answer.
Process timeline and approvals
Tool progress, skill activation cards, artifacts, MCP approval drawers, and human-in-the-loop prompts make agent work visible and resumable.
isa_agent_sdk runtime depth
The app uses SDK-grade primitives for browser observation, artifacts, connector/tool approvals, multi-step execution, and process orchestration.
Eight apps, one memory.
Distinct workspaces that share the same agent, account, and memory — switch without losing context.
Chat
Agentic chat with live tool use, extended thinking, and runnable code artifacts.
Dream
Generate and refine images and creative designs.
Hunt
Search across products, learning, community, and the open web.
Omni
Draft content across formats and topics from templates.
Knowledge
Ask your documents — vector + graph retrieval.
Data Scientist
Query, visualize, and model data from CSV, a database, or a URL.
Trading
A quant workspace: portfolio, alerts, screening, daily reports.
Maestro Workflows
Inspect live workflow templates and runs, then launch through the configured Maestro runtime.
60+ built-in tools.
Curated and loaded into every session — the agent picks what the task needs.
Files & code
Read, write, edit, search, and run shell commands.
Web & browser
Search the web and drive a real browser — observe, act, extract.
Tasks & calendar
Create tasks, set reminders, schedule calendar events.
Memory
Store and recall working memory across sessions.
Media & creative
Generate images, analyze visuals, transcribe audio, design brand assets.
Documents
Create documents, spreadsheets, presentations, and PDFs.
Meet users where they are.
Run the agent through the chat apps your team and customers already use.
Live: Telegram · Roadmap: WhatsApp, Feishu, Slack, Discord, Signal, Teams
One agent, or a team of them.
Delegate to a specialist, or fan a task across a swarm — sequential, streaming, or as a DAG.
Trade
Markets, portfolio, screening, alerts.
Creative
Brand, design, copywriting.
Marketing
Campaigns, analytics, SEO, growth.
Vibe
Code orchestration, testing, debugging, refactor.
Built on a real platform.
The production layer that makes agents controllable, observable, and auditable.
Model routing
75+ models across providers, with routing and failover policies for availability and task fit.
Persistent memory
Tiered memory with summarization and a knowledge graph.
Human in the loop
Approve, answer, or correct mid-run — with a full audit trail.
Bring your own keys
Register, rotate, and audit your own provider credentials.
Artifacts & sandbox
Live code execution and preview inside the chat.
Full tracing
Every run traced — inspect, replay, and compare.
isagent.io/console
The console makes agent infrastructure user-operable.
Users and platform owners can manipulate MCP servers, skills, tools, agents, multi-agent systems, orchestration graphs, workflows, models, runtime policy, and observability without treating the platform as a black box.
MCP servers and tools
Register servers, inspect available tools, review schemas, search tool catalogs, and test tool behavior from the console surface.
Skills and prompt assets
Browse and manage skills, skill categories, prompts, and tool-to-skill mappings so users can shape what agents are allowed to do.
Build and configure agents
Create agents, tune advanced configuration, connect tools and skills, manage schedules, and publish usable agent profiles.
Multi-agent orchestration
Manage multi-agent specs, agent graphs, Vibe-generated teams, routing, collaboration roles, and orchestration policies.
Workflow and DAG operations
Design DAGs, validate workflows, run executions, inspect execution detail, and operate repeatable processes rather than isolated prompts.
Model, runtime, and observability controls
Control providers, model routes, runtime resources, audit evidence, usage, traces, and operational health from the same admin plane.
Enterprise-ready foundations
SSO / SAMLRole-based accessAudit logsBring your own keysSelf-host or SaaS
See it work for you.
Start free and put every one of these to work in your first session.