Describe the job. Get AI that does it.
augLab is the AI agent platform for teams. Chat with 130+ tools, turn any task into an agent or workflow in plain English, and run it all with approvals, privacy filters and cost tracking, on the web or from your iPhone.
400+ models from the labs you already trust
One chat. 130+ tools. Real deliverables.
Ask for the brief, the deck, the analysis or the website, and watch it get made. Web search is built in, and the same conversation can build the agents that do it again tomorrow.
Your apps, inside the chat
Gamma, Asana, Duda, HeyGen and more connect right in the conversation, with every one of their tools in reach.
Big jobs in one message
Up to 100 tool calls in a single turn, so multi-step research and builds finish without you nudging them along.
A work panel for what it makes
Pages, charts, Markdown, CSV, JSON, code, images, audio, video and PDFs open beside the chat as they're created.
Projects that remember
Give a project its own instructions, knowledge, skills and default model. Memory carries context from chat to chat.
Share it, schedule it
Share with teammates or by a public link that expires. Schedule a chat to run again on a cadence and land the results.
Builds your stack for you
Ask for an agent, team, workflow, skill, knowledge base or website form, and the chat creates it for you to review.
Your AI team, in your pocket.
Not a companion app: the same chat, the same agents and the same inbox you use at your desk. Approve a refund from the checkout line. Rerun the morning briefing from the train.
- Full chat with model picker, tools, knowledge, skills and memory
- Attach files and photos, or snap one with the camera
- Create agents, teams and workflows by hand or with Create with AI
- Run, schedule and restore earlier versions on the go
- Approve or reject waiting runs from the Inbox
- Run history and cost analytics, behind a Face ID lock



Everything it takes to put AI to work
Build it, run it, watch it, and keep it safe, without stitching together five vendors.
Create with AI
Describe the job in plain English. AI designs a complete agent, team, or multi-step workflow, and you review every part before anything is created.
Visual Workflows
Branches, loops, routers, parallel steps and nested workflows on a canvas. Ask branch questions in plain English and watch test runs step by step.
Multi-Agent Teams
Specialist agents that collaborate on one job, with a leader deciding who handles what.
Decision Engine
Scores every run, screens tool results for prompt injection, and checks each agent's tools and model against its instructions.
Privacy Filters
Mask or block SSNs, card numbers, emails and phone numbers before they reach a model. Admins set a floor agents can't lower.
Cost Observability
Cost and tokens on every run and trace, spend by agent and model, monthly budgets, per-agent caps and alerts.
Knowledge Base (RAG)
PDFs, docs, spreadsheets and URLs, or synced Google Drive and OneDrive folders. Hybrid search with citations grounds every answer.
Inbox & Approvals
Approval gates before a run, a step or a tool call. Approve from one inbox on the web or your phone.
Versions & Publishing
Every save is a version. Compare, restore and publish; schedules, the API and widgets run the published version while you edit the draft.
Evals & Run Quality
Rubric test suites from the dashboard or CI, a quality score on every run, and side-by-side comparisons. LLM-judged evals on Business+.
Schedules & Webhooks
Cron, intervals and webhooks put agents, workflows and even chats on autopilot.
augLab on iPhone
Chat, build, run, approve and check costs from your phone. Android is coming soon.
Say what you need. Review what it builds.
Write a sentence about the job. augLab picks the model, the tools and the instructions, then hands you a draft to tune before anything goes live.
- Model, tools and instructions filled in from a brief
- Setup suggestions flag missing tools or a better-fitting model
- Chat with unsaved changes in a live preview pane
- Approval gates, memory, knowledge, skills and output schemas
- Version history with draft and publish
Automations you can watch work.
Lay out a multi-step job, ask branch questions in plain English, and watch a test run light up step by step before you publish. No code unless you want it.
- Drag steps from the palette, or tap + on any connector
- Agents, teams, Python functions and nested workflows as steps
- Conditions, routers and loops in plain English or Python
- Parallel groups, per-step timeouts and skip-on-failure
- Structured JSON output from any step
- Undo and redo, a guided tour, and a to-finish checklist
No surprise AI bills.
See what every run, agent and model costs as it happens, and set limits before anything runs away.
- Tokens and cost on every run, and on every step of a trace
- Spend by agent, team, workflow and model over 7, 30 or 90 days
- A monthly org budget, per-agent caps and alert thresholds
- Alerts for budgets, failed runs, waiting approvals and eval regressions
- Web search usage and paid searches on the Billing page
- Decision engine checks are on us, never billed to you
AI that checks its own work.
A fast judgment model and org-wide privacy filters run alongside every agent, so quality, safety and sensitive data aren't left to chance.
- A quality score on every completed run, judged against its task
- Web, email and CRM results screened for prompt injection
- Setup checks catch the wrong tools or model before you save
- SSNs, card numbers, emails and phone numbers masked or blocked
- Admins set the policy; agents can tighten it, never loosen it
- Each check can be off, measure-only or on, and judged text is never stored
Software, plus people who've done this before
We don't just hand you a login. We partner with you to deploy AI that delivers results.
Onboarding
We learn your business, find the high-impact use cases, and design your first agents together.
Build & Configure
Our team helps configure agents, workflows and integrations tailored to how you work.
Train & Test
Your team learns the platform while we tune agents for accuracy and reliability.
Deploy & Scale
Go live with confidence, with ongoing support, monitoring and optimization as you grow.
Built for every team
From support to sales, research to reporting, augLab agents take the busywork so your people can do the work that matters.
Works where your team already works
130+ built-in tools, plus any HTTP API or MCP server you add. Just added: BigQuery, Redshift, Neo4j, Browserbase, Apify, Webex, Amazon SES, AWS Lambda and more.
Studio and Team include a curated set of tools. Business and Enterprise get the full catalog with your own keys.
Put your best agent on your website
Chat widgets and AI-powered forms, connected to any agent, team or workflow. Branded, domain-restricted and live with one script tag.
- Chat widgets powered by your agents
- Forms that start runs or feed a knowledge base
- Describe the widget you want and AI sets it up
- Custom branding, theme and domain restrictions
- Or give the agent a face and voice with augLab Avatars
Runs wherever the work happens
REST API
Run agents, teams and workflows, continue sessions, run async with signed webhooks, and trigger evals from CI.
MCP & A2A
Manage augLab from any MCP client, or publish an agent, team or workflow as its own MCP or A2A endpoint. Team plan and up.
Schedules & Webhooks
Cron schedules, intervals and webhooks, with timezones and a Run Now button.
Website
Chat widgets, forms and talking avatars with a single script tag.
iPhone
Chat, run, approve and track costs from the augLab app.
The augLab ecosystem
Five products, one login. Your agents can now work across all of them.
Ready for your security review
Your data stays yours. We don't train models on it.
Your Own Keys
Business and Enterprise use your own provider keys with no token markup.
Your Models, Your Cloud
Run models you host in Azure OpenAI, AWS Bedrock, Google Vertex AI or any OpenAI-compatible endpoint. Business and Enterprise.
Privacy Filters
SSNs and card numbers masked by default; emails and phone numbers too if you choose. Org-wide policy with an enforced minimum.
Encrypted at Rest
Credentials, chats and memories encrypted field by field with AES-256. Logs redact personal identifiers and secrets.
HIPAA Mode
On Enterprise: approved-provider enforcement, required MFA and audit logging. BAA by arrangement.
Roles & Access
Viewer, runner, member, editor and admin roles, plus an option to lock shared items to their creators.
Audit Trail & SSO
Hash-chained audit logs, SAML and OIDC single sign-on, and enforceable MFA.
No Vendor Lock-in
Export any agent, team or workflow as a bundle. Standard APIs. Switch models anytime.
Quick answers
What is augLab?
augLab is an AI agent platform for teams. You chat with AI that can use 130+ tools and connected apps, then turn that work into agents, multi-agent teams and workflows that run on schedules, webhooks, your website, the API, or from the augLab iPhone app.
Is there an augLab mobile app?
Yes. augLab for iPhone is on the App Store, and Android is coming soon. The app includes the full chat, building and editing agents, teams and workflows, running and scheduling them, approving requests from the Inbox, and run history with cost analytics.
How does augLab protect sensitive data?
Privacy filters mask or block Social Security numbers, card numbers, email addresses and phone numbers in prompts and mask them in replies. Org admins set the policy and agents can only tighten it. Sensitive fields are encrypted at rest with AES-256, logs redact personal identifiers, and Enterprise can turn on HIPAA mode.
How do I see and control what AI costs?
Every run and trace shows its tokens and cost. Analytics breaks spend down by agent, team, workflow and model over 7, 30 or 90 days. Admins set a monthly budget, per-agent caps and alert thresholds, and the decision engine's checks are never billed to you.
Which AI models does augLab support?
400+ models from 11 providers, including OpenAI, Anthropic, Google, Mistral, xAI, DeepSeek and Perplexity. On Business and Enterprise you can bring your own keys or run models you host in Azure OpenAI, AWS Bedrock, Google Vertex AI or any OpenAI-compatible endpoint.
Start in minutes. Scale without surprises.
Self-serve Studio and Team with a 3-day trial, prepaid usage and web search included. Business adds your own keys and invoicing.
Studio
1 seat · $20 credit · 3-day trial
Self-serve. Platform-managed models and a curated toolset. 1,000 web searches a month.
Start StudioTeam
5 seats · $100 credit · MCP
Higher limits, MCP and $49 extra seats. 5,000 web searches a month. Platform-managed models.
Start TeamBusiness
From · Invoice · BYOK
Unlimited usage, your own keys and models, the full tool catalog, and Login with augLab across the ecosystem.
Learn MorePut AI to work this week.
Book a demo and we'll build your first agent with you, or start a trial and build it yourself.