Axiom.ai
No code browser automation tool and Chrome extension for building bots that scrape, fill forms, and move data across websites, buildable by clicking, coding, or describing to Claude.
axiom.ai is a browser automation and web scraping tool for building bots that click, type, scrape, fill forms and move data across websites. It offers three ways to build the same bot. In the no code builder in the Chrome extension, a bot is placed step by step or described in plain English and assembled by AI. Developers can code against axiom.ai's cloud browser, and a Claude skill builds a no code or code bot from a description inside Claude Code. Its templates cover separate jobs such as automating Apollo, Facebook posts and file transfers.
Bots chain steps with loops, logic branches and error handling, and an Integrate AI step passes data mid flow to the LLM the customer picks, including ChatGPT, Claude, Gemini and Sonar, on the customer's own API key.
They connect to Zapier, Make, n8n, Google Sheets, Excel, Airtable and Retool, and run manually, on a schedule, from a webhook or through the Orchestrate REST API, in axiom.ai's cloud or on the customer's own machine through the desktop app, which keeps login data on the customer's network.
Developers can also drive the cloud browser over CDP with Puppeteer or Playwright, use the Step API Node library, or connect AI agents through the MCP server in the desktop app. The dashboard keeps run reports, and runs can be recorded for debugging.
Pricing is by monthly runtime hours, from Starter at $15 for 5 hours to Ultimate at $250 for 250 hours, after 2 free hours for every new account. axiom.ai publishes no roles, SSO or security attestations, keeps no memory between runs and offers no way to test bots against scored cases, and its bots follow the steps they are built with rather than planning their own.
From our comparison video
Axiom.ai vs Bardeen: Runtime Hours or Credits per Row
Our 4 minute video puts Axiom.ai side by side with Bardeen. It covers what each one does and how each is priced, then when to choose which.
Recorded in October 2026. The profile here is live and updates as research is verified.
Vendor details
Canonical URL
https://axiom.ai
Category
Browser / computer-use agent
Company status
independent
Use cases & customers
In practice
A marketer needs competitor prices tracked daily. In Axiom they point and click to build a scraper on the target site, schedule it to run every morning in the cloud, and pipe the extracted data straight into Google Sheets.
A developer wants browser automation inside their own stack. They connect Puppeteer to Axiom's cloud browser over the Chrome DevTools Protocol, running headless browsers on Axiom's infrastructure straight from their code without managing servers.
An operations team faces hours of repetitive form filling. They record the workflow once as a no code bot, add error handling, and let it submit entries across a web application on a schedule, freeing staff for higher value work.
Sources & related URLs
Research sources
Agentic Index coverage score
10.0 / 14 capabilities · 71%
| Integrations & Tool Calling | Full |
|---|---|
|
An integrations catalog connects bots to Zapier, Make, n8n, Google Sheets, Excel, Airtable, Retool, ChatGPT, Claude, Gemini, Sonar and 2Captcha, reading and writing data, with webhooks for other systems. It also covers DeepSeek, 1Password and Relay.app. Webhooks, API access and the Zapier and Make connections start on the Pro plan. SourceAxiom.ai, axiom.ai/llms.txt and /pricingread 2026-10-05 |
|
| Workflow Orchestration | Full |
|
Bots chain steps with loops, logic branches, error handling settings and Integrate AI steps that pass data to an LLM mid flow, run concurrently or across machines, and are triggered and managed through the Orchestrate API. Builders can also add data transformation and validation steps, work across several tabs and move files in and out. In the builder a bot can be placed step by step or described in plain English, and the AI assembles the steps, which stay editable. The Claude skill does the same from Claude Code, building a no code or code bot. In the cloud, Pro Max runs 2 bots at once and Ultimate runs 20. SourceAxiom.ai, axiom.ai home, /pricing and llms.txtread 2026-10-23 |
|
| Knowledge Grounding & RAG | Not documented |
|
Bots import data from sheets and apps and pass it to an LLM step at run time, and axiom.ai keeps no store of the customer's documents for a bot to search. The Integrate AI step works only on the data passed into it: it pulls named fields out of scraped text into columns, or generates text from a prompt that mixes in data from earlier steps. SourceAxiom.ai, axiom.ai/docs Integrate AI and llms.txtread 2026-10-05 |
|
| Human Oversight & Guardrails | Partial |
|
A Pause step stops a bot so a person can interact with the page before it continues, and scheduled bots can be paused or stopped from the dashboard. axiom.ai offers no approval step that holds an action until someone signs off. Runs started through the Orchestrate API can also be stopped while they are in flight. SourceAxiom.ai, axiom.ai/docs Wait and Pause steps and Orchestrate APIread 2026-10-05 |
|
| Security, Identity & Governance | Not documented |
|
Team accounts share runtime and billing, and the desktop app keeps login data on a machine the customer controls. axiom.ai publishes no roles, SSO or security attestation, and axiom.ai/security returns 404. On the desktop app, none of the processed data leaves the machine; only account information and run metadata such as start time and duration go to axiom.ai. Session data is encrypted at rest and in transit, and team accounts with member invitations come on the Ultimate plan. SourceAxiom.ai, axiom.ai/docs secure login and /pricingread 2026-10-05 |
|
| Observability & Auditability | Full |
|
The dashboard keeps run reports for each automation, runs can be recorded for debugging, and a debugger and error metadata steps show step results. There is no export to a SIEM, and no retention period is given. Run recording starts on the Pro plan. Through the Orchestrate API, a developer can check a run's status and fetch its run history. SourceAxiom.ai, axiom.ai/docs dashboard and Orchestrate API, and /pricingread 2026-10-05 |
|
| Memory & State Persistence | Not documented |
|
Saved site logins, cookies and Chrome profiles keep sessions across runs, and data a bot needs is imported for each run. Those keep a browser signed in, and axiom.ai describes no memory that a bot reads from one run to the next. On the desktop app, the browser cookie is shared automatically so the bot stays signed in, and all cookies are double encrypted. SourceAxiom.ai, axiom.ai/pricing and axiom.ai/docs secure loginread 2026-10-05 |
|
| Deployment & Data Residency | Full |
|
Bots run on axiom.ai's cloud or on the customer's own machine through the desktop app, which keeps login and session data on the customer's network; no region list is published. The desktop app sends axiom.ai only account information and run metadata, and it has its own scheduler, webhook triggers and MCP server. Each cloud run is capped by plan, at 1 hour on Starter, 2 hours on Pro, 3 hours on Pro Max and 12 hours on Ultimate, while desktop runs have no concurrency limit. Bots can also run on a VPS. SourceAxiom.ai, axiom.ai/docs desktop app and secure login, and /pricingread 2026-10-05 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
axiom.ai publishes templates for separate jobs (automating Apollo, Facebook posts, file downloads and uploads and others) with video guides, and bots can be shared as templates. The Apollo template adds prospects to lists in Apollo.io, and its video guide walks through the setup. SourceAxiom.ai, axiom.ai/video-guides/templates/automate-apolloread 2026-10-05 |
|
| Triggers & Channel Coverage | Full |
|
Bots run manually, on a schedule in the cloud or through the desktop app's built in scheduler, or when triggered by a webhook, Zapier, Make or the Orchestrate REST API. Cloud schedules run daily on Pro, hourly on Pro Max and every 15 minutes on Ultimate, while the desktop scheduler can run a bot every minute. Make can also start a run on the desktop app. SourceAxiom.ai, axiom.ai/llms.txt and /pricingread 2026-10-05 |
|
| Model Flexibility & Routing | Full |
|
The Integrate AI step sends data to the LLM the customer picks, with ChatGPT, Claude, Gemini and Sonar available as integrations. DeepSeek is listed too, and the customer supplies an API key for the provider it chooses. The step either extracts a list of named fields into columns or generates free text from a prompt. SourceAxiom.ai, axiom.ai/docs Integrate AI and llms.txtread 2026-10-05 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
Developers get three APIs, the Chrome API over CDP at wss://cdp-lb.axiom.ai, the Step API Node library @axiom_ai/api, and the Orchestrate REST API for triggering and managing saved automations, plus an MCP server and a Claude skill. The Chrome API takes an API key in the WebSocket address and opens a fresh cloud browser for any CDP client. The Step API wraps actions such as opening a browser, going to a page, clicking, typing and scraping. The Orchestrate API starts any saved automation, checks and lists runs and stops them, at up to 100 calls a minute per key. The MCP server ships inside the desktop app and exposes automations as tools for Claude Desktop, Claude Code and other MCP clients. SourceAxiom.ai, axiom.ai/docs developer hub API and llms.txtread 2026-10-05 |
|
| Testing, Debugging & Optimization | Partial |
|
A debugger, error metadata steps and run recordings help find why a run failed, and error handling is configured for each bot. axiom.ai publishes no test cases or scores for checking a bot's results and no check a bot must pass before it runs. The AI Assistant in the builder suggests fixes while debugging. SourceAxiom.ai, axiom.ai/docs troubleshooting and llms.txtread 2026-10-05 |
|
| Browser & Computer Use | Full |
|
axiom.ai bots drive a real Chrome browser, clicking, typing, navigating, handling logins and scraping on any site, from the Chrome extension, the desktop app or axiom.ai's cloud browser, which developers can also drive over CDP with Puppeteer or Playwright. Bots sign in with saved logins or 1Password and pass two factor codes, manage several tabs, and upload and download files. A Turnstile bypass gets past Cloudflare CAPTCHA challenges on the desktop app on every plan and in the cloud on Ultimate, and the 2Captcha integration solves other CAPTCHAs. Customers can bring their own rotating proxies. SourceAxiom.ai, axiom.ai/llms.txt and /pricingread 2026-10-23 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
From $15 a month (Starter, 5 runtime hours) · 2 free hours to try
runtime hours a month
Included quota
Starter: 5 runtime hours a month, 1 bot at a time in the cloud with a 1 hour cap per run, no code and code tools, 2FA logins. Pro: 30 hours, webhooks and the API, daily cloud schedules, Zapier and Make, run recording and the desktop MCP server, with a 2 hour cloud cap. Pro Max: 100 hours, 2 cloud bots at once, hourly cloud schedules, a 3 hour cap and a Slack channel. Ultimate: 250 hours, 20 cloud bots at once, cloud schedules every 15 minutes, a 12 hour cap, Turnstile bypass in the cloud and team accounts. Desktop runs have no concurrency or time limit on any plan.
What is public
axiom.ai publishes all four plans with prices, runtime hours, cloud concurrency, per run caps, schedule frequency and support level, plus a feature table by plan.
Billing mechanics
Self serve monthly subscription priced by runtime hours, with four plans from $15 to $250 a month. Runtime counts on the desktop app and in the cloud, and every new account starts with 2 free hours and no card.
Cost watchouts
LLM steps run on the customer's own API key, so model usage is billed by the provider on top of the plan. Cloud runs stop at the plan's per run cap (1 hour on Starter up to 12 on Ultimate), and webhooks, the API and cloud schedules need Pro or above.
Variable cost rationale
Flat monthly plans sized by runtime hours; heavier use moves a customer up a plan, and only the customer's own LLM usage is metered separately.
Additional watchouts
The model cost of AI steps sits outside the plan, cloud concurrency is 1 bot until Pro Max, and team accounts come only on Ultimate.
Overage / add-ons
No overage price is published; more runtime means a larger plan.
Free / trial
2 free hours of runtime for new accounts, no credit card
Lowest paid plan
Starter at $15 per month with 5 hours of runtime
Key ambiguities
The Ultimate card says 20 cloud bots at once, while a tooltip in the feature table says Ultimate users can run up to 30; the card figure is used here.
Support SLA / resale
Starter has email and community support, and Pro email support. Pro Max adds a dedicated Slack channel, and Ultimate priority Slack and email support.
Related vendors
- AgentQL — A self healing query language for the web, with natural language…
- AGI, Inc. — AGI, Inc
- Airtop — Cloud browsers for AI agents that compile plain English workflows…
- Asteroid — Healthcare portal integration platform whose supervised browser and…
- Automat AI — AI agents that operate computers visually, sold as Automat Workforce…
- Autotab — General AI agent that learns a workflow from a demonstration and…
Alternatives to Axiom.ai
The closest documented capability profiles to Axiom.ai among browser and computer-use agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Airtop9.0 / 14Adds documented Security, Identity & Governance
- H Platform9.0 / 14Adds documented Security, Identity & Governance
- Browser Use11.5 / 14Adds documented Security, Identity & Governance and Memory & State Persistence
- Notte9.5 / 14Adds documented Security, Identity & Governance
- Skyvern11.5 / 14Adds documented Security, Identity & Governance and Memory & State Persistence
- Autotab8.0 / 14Adds documented Memory & State Persistence
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded