Releases: getmaxun/maxun
v0.0.38 | 9.7x Faster Extraction + Smart Queries
β‘ Performance Improvements
Maxun is now significantly faster across core workflows
- Extract β 9.74Γ faster
- Scrape β 5.71Γ faster
- Crawl β 5.31Γ faster
- Search β 2.09Γ faster
β¨ Smart Queries
You can now attach natural language instructions to your scraping workflows.
With Smart Queries, Maxun doesnβt just scrape pages β it can understand and extract exactly what you need using AI.
How it works
Add a prompt while creating a scrape robot. After the page is scraped, an LLM analyzes the content and returns a structured answer β no extra setup required.
Examples
- βList all pricing plans and their pricesβ
- βFind the latest blog post title and dateβ
Output
Runs now include a new field
{
"markdown": "...",
"html": "...",
"promptResult": "The pricing plans are: Starter ($9/mo), Growth ($29/mo), Pro ($99/mo)."
}Works Via
- Dashboard (No-Code)
- API
- SDK
- CLI
Learn more
https://docs.maxun.dev/robot/scrape
What's Changed
- fix: filter out separators by @RohitR311 in #1034
- fix: enable pagination highlighting by @RohitR311 in #1035
- fix: improve markdown conversion by @RohitR311 in #1037
- fix: malformed URLs not being processed by @RohitR311 in #1038
- feat: ssl config for hosted db by @RohitR311 in #1039
- feat(core): speed up robot runs by @RohitR311 in #1040
- fix: scrape robot edit section by @RohitR311 in #1043
- feat: add text content support for scrape by @RohitR311 in #1041
- feat: post scraping smart queries by @RohitR311 in #1042
- chore(deps): upgrade vite version to 6.4.2 by @amhsirak in #1036
- chore: pre-release v0.0.38 by @amhsirak in #1044
Full Changelog: v0.0.37...v0.0.38
v0.0.37
What's New
π¦ OpenClaw & Claude Skill Support
Maxun now supports OpenClaw and Claude Code skills.
You can instruct your AI directly - βlist my robots,β βrun this robot,β βfetch the latest run data,β etc.
No API calls needed - the AI performs all Maxun actions for you automatically.
Learn more:
- Openclaw: https://docs.maxun.dev/integrations/openclaw
- Claude: https://docs.maxun.dev/integrations/claude-code
What's Changed
- feat: Added Output Format Selection to Crawl and Search Robots by @Rutetid in #1009
- fix: success status being displayed by @amhsirak in #1017
- feat: add cli to main menu by @amhsirak in #1019
- feat: add run status for cli and mcp by @RohitR311 in #1022
- fix: copy clipboard hosted context by @RohitR311 in #1024
- feat: use dialog instead of custom generic modal by @amhsirak in #1020
- fix: add missing i18n translation keys across all locales by @Smeet23 in #1025
- feat: add claude code integration by @amhsirak in #1028
- fix: misc ui fixes by @amhsirak in #1029
- chore: pre-release v0.0.37 by @amhsirak in #1030
New Contributors
Full Changelog: v0.0.36...v0.0.37
v0.0.36
What's New
π Robot Duplication via SDK
You can now duplicate any robot through the SDK, the same way you already can through the UI and API.
π¦ Maxun CLI
Weβre releasing the official Maxun CLI, available for both Maxun Cloud and Open Source users.
The CLI lets you create, manage, run data scraping robots. It's a simple way to work with Maxun from your terminal.
Learn more: https://docs.maxun.dev/category/cli
Repository: https://github.com/getmaxun/maxun-cli
Maxun_CLI.mp4
What's Changed
- fix: add langgraph svg by @RohitR311 in #1011
- feat: add status check route sdk for Maxun CLI by @RohitR311 in #1012
- fix(recorder): auto-discard recording session after 10-minute timeout by @ritam1303 in #999
- fix: enforce unique robot names across all creation flows and edit by @ritam1303 in #998
- feat: add sdk robot duplication by @RohitR311 in #1005
- fix: error handling apikey check by @RohitR311 in #1010
- fix: sdk doc link path by @RohitR311 in #1014
- fix: duplicate field naming for captured data by @RohitR311 in #1015
- chore: pre-release v0.0.36 by @amhsirak in #1016
New Contributors
- @ritam1303 made their first contribution in #999
Full Changelog: v0.0.35...v0.0.36
v0.0.35
What's New
π Robot Duplication
You can now duplicate Extract Robots created using Recorder mode to quickly create a new robot from an existing configuration.
This makes it easy to reuse extraction workflows or run similar automations without rebuilding the robot from scratch.
Robot duplication is available from the UI and can also be done programmatically via the API.
Learn more: https://docs.maxun.dev/robot/extract/robot-duplicate
What's Changed
- Fix: Enable recorder browser reload with proper navigation handling (#988) by @alwaysalearner1234 in #997
- feat: add robot duplication by @RohitR311 in #993
- fix: robot edit url logic by @RohitR311 in #994
- fix: generate uuid for non secure context by @RohitR311 in #1000
- feat: remove standard proxy tabs by @amhsirak in #1001
- feat: improve realtime scrolling action by @RohitR311 in #1002
- chore: pre-release v0.0.35 by @amhsirak in #1003
New Contributors
- @alwaysalearner1234 made their first contribution in #997
Full Changelog: v0.0.34...v0.0.35
v0.0.34 | Maxun Python SDK
Weβre excited to announce the release of the Maxun Python SDK v0.0.1!
Maxun already offers a Node.js SDK for building web automation and data extraction workflows β and now the same capabilities are available for Python developers as well.
You can now programmatically create, run, and manage robots directly from your Python applications.
Documentation: https://docs.maxun.dev/category/python-sdk
Repository: https://github.com/getmaxun/python-sdk
What's Changed
- fix: api key modal close by @amhsirak in #982
- feat: enhance auto pagination detection logic by @RohitR311 in #976
- fix: auto robots button ui by @RohitR311 in #983
- chore: cleanup by @amhsirak in #985
- fix: integration alt text by @amhsirak in #986
- fix(core): normalize scrape schema by @amhsirak in #987
- chore: pre-release v0.0.34 by @amhsirak in #991
Full Changelog: v0.0.33...v0.0.34
v0.0.33 | Bug Fix Release
What's Changed
- fix: highlighting ui by @RohitR311 in #961
- fix: broken crawl docs link by @amhsirak in #962
- chore(ts): set
skipLibChecktrue by @amhsirak in #960 - feat: confirm before api key deletion by @amhsirak in #963
- fix(core): scroll down pagination by @RohitR311 in #964
- feat: revamp export for robot run data by @RohitR311 in #965
- fix: increase mem allocation to 4gb by @RohitR311 in #966
- fix: grouping elements with no attributes by @RohitR311 in #969
- fix: rm recorder script injection by @RohitR311 in #970
- fix: register navigation action by @RohitR311 in #971
- fix: match nav items theme by @amhsirak in #972
- fix: remote browser scroll by @RohitR311 in #973
- feat: auto detect pagination ui by @amhsirak in #974
- chore: pre-release v0.0.33 by @amhsirak in #975
Full Changelog: v0.0.32...v0.0.33
v0.0.32 | AI-Native Workflows & Real-Time Recorder
Whatβs New
π€ LLM Integrations
You can now plug Maxun directly into popular AI frameworks using our SDK. This makes it much easier to build AI-powered workflows on top of Maxunβs data layer.
Learn More: https://docs.maxun.dev/category/integrations
π§ Smarter AI Mode Extract (No URL Required)
AI Mode Extract no longer requires a website URL upfront.
You can now use natural prompts like
βExtract 10 companies from the Y Combinator websiteβ
Maxun will automatically identify the source and perform the extraction for you.
AI_Mode_Extract_No_Url.mp4
π₯ Enhanced Recorder with Real-Time Browser Sync
Recorder mode now stays in real-time sync with the live website, making recordings more accurate and reliable.
Youβll see and capture actual browser actions as they happen, including
- Typing into inputs
- Clicking buttons and links
- Scrolling pages
- Interacting with dynamic elements
What's Changed
- chore: deprecate robot duplication by @amhsirak in #941
- fix: recorder flow by @RohitR311 in #942
- fix: crawl and search data rendering by @RohitR311 in #945
- feat: add auto search logic by @RohitR311 in #946
- fix(core): max depth crawling logic by @RohitR311 in #947
- feat: add auto robots template by @RohitR311 in #943
- fix: swagger api routes display by @RohitR311 in #948
- feat: sync theme w cloud by @amhsirak in #949
- feat: rename name & target url labels by @amhsirak in #950
- feat(ui): redesign main menu layout and navigation by @amhsirak in #951
- fix: modify api key generation by @RohitR311 in #952
- feat: llm integrations by @amhsirak in #953
- fix: alert cloud sync by @amhsirak in #955
- feat: upgrade redirect to docs by @amhsirak in #956
- chore: pre-release v0.0.32 by @amhsirak in #954
- Revert "feat: add auto search logic" by @amhsirak in #957
- Revert "feat: add auto robots template" by @amhsirak in #958
- feat: browse auto robots by @amhsirak in #959
Full Changelog: v0.0.31...v0.0.32
v0.0.31 | Autonomous Web Discovery & Search
v0.0.31 introduces two powerful new robots to the Maxun engine: Crawl and Search. These features allow you to automate data discovery at scale, whether you are mapping entire domains or researching the web via natural language.
πΈοΈCrawl
Stop manually feeding URLs. Crawl intelligently discovers and extracts entire websites.
- Intelligent Discovery: Uses both Sitemap parsing and Link following to find every relevant page.
- Granular Scope Control: Target exactly what you need with Domain, Subdomain, or Path-specific modes.
- Advanced Filtering: Use Regex patterns to include or exclude specific content (e.g., skip
/admin, target/blog/*). - Depth Control: Define how many levels deep the robot should navigate from your starting URL.
Deep_Crawl.mp4
π Search
Turn search engine queries into structured datasets.
- Query Based: Search the web with a search query - same as you would type in a search engine.
- Dual Modes: Use Discover Mode for fast metadata/URL harvesting, or Scrape Mode to automatically visit and extract full content from every search result.
- Recency Filters: Narrow down data by time (Day, Week, Month, Year) to find the freshest content.
Deep_Search.mp4
Hybrid Access
Manage these robots through our no-code dashboard or trigger them programmatically via our Developer SDK.
What's Changed
- fix: remove
onCancelbutton from robot config page by @amhsirak in #929 - feat: cleaner alert notifs by @amhsirak in #927
- feat(ui): match theme of standard & outlined alerts by @amhsirak in #928
- feat: !show rows per page for runs & robots by @amhsirak in #930
- fix: show formats selected for scrape by @amhsirak in #931
- chore: pass prompt by @RohitR311 in #932
- feat: auto name lists for ai mode by @RohitR311 in #933
- feat: !wrap robots table w table head by @amhsirak in #934
- feat: display api key creation date by @amhsirak in #935
- feat: run scroll customization by @amhsirak in #936
- feat(core): real time progress update for runs by @amhsirak in #937
- feat: add crawl and search by @RohitR311 in #938
- chore: pre-release v0.0.31 by @amhsirak in #940
Full Changelog: v0.0.30...v0.0.31
v0.0.30 | π₯Maxun SDK + AI Mode
What's New
π₯ AI Mode For Extraction (Beta)
Now build Extract Robots without recording! Simply describe what you want in plain language.
Example: Extract Names, Rating & Duration of Top 50 Movies from IMDb
maxun_ai_mode_extraction.mp4
Choose your approach
- Recorder Mode: Click through websites to record actions step-by-step.
- AI Mode: Describe what you want - the AI handles the extraction logic.
Supports local LLMs via Ollama for privacy and cloud models (OpenAI, Anthropic) for maximum accuracy.
Get Started with AI Mode: https://docs.maxun.dev/robot/extract/llm-extraction
π¦ Maxun SDK v0.0.1
The official Node.js SDK for Maxun.
Works with both Maxun Cloud and Maxun Open Source - automatically handles the differences for you.
Get Started: https://github.com/getmaxun/node-sdk
Core Features
1. LLM Extraction (Beta) β Extract data using natural language prompts
Example: Extract the first 15 Company Names, Descriptions and Batch Information from YC
maxun_sdk_llm_extract.mp4
2. Non-LLM Extraction β Use precise selectors for reliability
a. Auto Pagination
Pagination is completely optional. When you don't specify the pagination field, Maxun automatically detects and handles pagination for you.
.captureList({
selector: '.product-card',
maxItems: 100
})b. Auto List Capture
When using captureList, you only need to provide the list item selector. Maxun automatically:
Detects all meaningful fields within each list item
Extracts clean, structured data from those fields
.captureList({
selector: '.product-card' // That's it! Maxun finds all fields inside
})Complete Robot Management
- Scheduling β Set up recurring runs with cron syntax
- Webhooks β Receive notifications on job completion
- Runs & Execution β Monitor status, results, and history
- Robot CRUD β Create, update, and manage robots programmatically
View all examples here: https://docs.maxun.dev/category/sdk
What's Changed
- feat: support for sdk by @RohitR311 in #920
- feat: AI powered robot generation by @RohitR311 in #921
- fix: set showCancelButton to false for schedule settings by @amhsirak in #922
- fix: misc ui & theme changes by @amhsirak in #923
- chore: pre-release v0.0.30 by @amhsirak in #924
Full Changelog: v0.0.29...v0.0.30
v0.0.29 | Screenshot Support for Scrape Robots!
What's New
β‘ Screenshot Support for Scrape Robots!
Scrape robots now support screenshots β just like Extract Robots.
Every run can now return
- Clean HTML
- Clean Markdown
- Full-page screenshots
- Visible-viewport screenshots
A more complete, visual snapshot of any page β without extra setup.
Scrape_Screenshots.mp4
What's Changed
- fix: multi-platform docker build by @RohitR311 in #913
- fix: user id error handling by @RohitR311 in #914
- fix: auto extract all fields by @amhsirak in #915
- feat: add scrape screenshot support by @RohitR311 in #916
- fix: accurate similar container grouping by @amhsirak in #909
- chore: pre-release v0.0.29 by @amhsirak in #917
Full Changelog: v0.0.28...v0.0.29