Tsenta vs AIHawk, an open-source auto job applier vs a managed agent

AIHawk (Auto_Jobs_Applier_AIHawk) was the best-known auto job applier on GitHub: a free Python script that signed in to your LinkedIn account and filled Easy Apply forms with an LLM. As of October 30, 2026, that GitHub link redirects to a different project by the same author, and the LinkedIn automation came out in 2025, so if you want an open-source LinkedIn applier that's still maintained, look at GodsScion's Auto_job_applier_linkedIn.

Tsenta is the managed option. it's an AI job application agent: it finds roles on 50,000+ company career pages, tailors your résumé for each one, and applies on the employer's own system across 30+ ATSes, from its cloud. no Python, no config files, and your LinkedIn account isn't part of it.

this isn't a takedown. the scripts are free, you can read every line, and they run on your own machine. for some people that's the right trade. below is what each tool does today (we checked both repos on October 30, 2026), what it costs, and who it suits.

what happened to AIHawk on GitHub

short version: the AIHawk you've read about is no longer what's at that link.

• 2024. Auto_Jobs_Applier_AIHawk opened Chrome, logged in to LinkedIn with your credentials, searched with the filters in your config file, clicked Easy Apply, and answered screening questions with an LLM (OpenAI, Claude, Gemini, or a local Ollama model). it could also generate a tailored résumé PDF for each application.
• by March 2025. the README read: "LinkedIn requested the removal of all links to the platform and banned all project contributors, so to comply with their guidelines, we have removed any form of automation on LinkedIn." what stayed in the repo was a résumé and cover letter generator.
• October 2025. the README said third-party provider plugins had been removed "due to copyright considerations."
• April 16, 2026. the repository was archived and went read-only.
• today. github.com/feder-cr/Auto_Jobs_Applier_AIHawk redirects to feder-cr/invisible_dots, which describes itself as an "Open-source, self-hosted alternative to OpenAI Dots, Grok Bot." it's MIT-licensed, its commit history starts on September 29, 2026, and it shows about 32,300 stars. most of those came from the job applier: the repo had about 29,000 in October 2025.

so if you search "ai hawk github auto apply" today, the LinkedIn applier you find is a fork of the 2024 code. one widely linked copy, 11844/Auto_Jobs_Applier_AIHawk, is a snapshot from October 29, 2024 with no commits since. a fork like that may still run, but nobody upstream fixes it when LinkedIn changes its pages.

the license moved around too. the 2024 code shipped with a custom source-available license that bars resale and commercial use without the author's permission (its README said MIT, so read the LICENSE file, not the badge). later versions were AGPL-3.0. the new project is MIT.

what GodsScion's Auto_job_applier_linkedIn does

this is the open-source LinkedIn applier that's still moving. as of October 30, 2026: about 3,000 stars, not archived, last commit on September 10, 2026, latest release v26.08.31, MIT license (releases before August 2026 were AGPL-3.0). one independent developer maintains it.

what it automates:

• searches LinkedIn with your filters, opens each matching job, and fills the Easy Apply form from your saved answers
• optionally drafts answers with an LLM. AI is off by default, and it supports OpenAI, Gemini, DeepSeek, or a local model through Ollama or LM Studio
• skips jobs by rules you set, including a filter for postings that say they don't sponsor visas
• for jobs that send you to the company's own site, it saves the external link to a CSV so you can apply by hand
• logs applied and failed jobs to CSV files, with a local history page

what's good about it: setup is friendlier than AIHawk ever was. you install Python 3.10+ and Chrome, double-click a launcher, and fill in tabs in a control panel that opens in your browser at a local address like 127.0.0.1:5000. there's a dry-run setting that fills every form and stops before submit, so you can see what it would send. your details stay in local files on your machine, unless you turn on a hosted AI provider, which then sees whatever you send it.

what to know: it's LinkedIn Easy Apply only. the Chrome window has to stay in the foreground unless you switch on background mode, which turns off the safety pauses. and its own docs say "LinkedIn caps Easy Apply at roughly 25 applications a day," so the README's "100+ jobs in under an hour" runs into a daily ceiling no script gets around.

setup: script vs managed agent

AIHawk (2024 fork). install Python and Chrome, clone the repo, create a virtual environment, install the requirements, then fill in three YAML files: secrets.yaml (your LLM API key), config.yaml (your search) and plain_text_resume.yaml (your résumé as structured data). your LinkedIn account language has to be English for it to work.

Auto_job_applier_linkedIn. install Python 3.10+ and Chrome, download the project, double-click the launcher for your OS, and fill in the control panel. you can still edit the config files by hand if you prefer.

Tsenta. create an account on the web or the iOS and Android apps, upload your résumé, and set your preferences. the agent runs in Tsenta's cloud, so nothing has to stay open on your computer.

ATS coverage: LinkedIn Easy Apply vs the employer's own system

both scripts work inside LinkedIn Easy Apply. that's a real slice of jobs. but plenty of LinkedIn postings show "Apply" instead of "Easy Apply," and LinkedIn's help center says those route you "to that company's website or job board to continue the application process." that's usually an ATS like Workday, Greenhouse, Lever, Ashby or iCIMS, and neither script fills those forms.

Tsenta works on the other side of that redirect. it finds roles on company career pages and submits on the employer's system: it handles the login or sign-up, fills every field, drafts the open-ended answers, uploads your documents, and submits. it supports 30+ ATSes, including Workday, Greenhouse, Lever, Ashby, iCIMS, SmartRecruiters and Oracle Cloud (full list).

the flip side is real too. Tsenta doesn't submit LinkedIn Easy Apply forms, so a role that only exists as an Easy Apply posting isn't one it can send.

LinkedIn terms of service exposure

LinkedIn's User Agreement (section 8.2, effective November 3, 2025) says members agree not to "Use bots or other unauthorized automated methods to access the Services." its help page on prohibited software says members who use tools that automate activity on LinkedIn "risk having their accounts restricted or shut down."

LinkedIn also sets daily and speed limits on Easy Apply. the daily count resets at the end of the day (UTC), rapid submissions trigger a temporary pause, and LinkedIn says "we may lower a member's daily Easy Apply limit if we notice activity that seems inauthentic."

both scripts drive your own LinkedIn account in a real Chrome window, so that exposure is yours. AIHawk is the cautionary tale: by its own README, LinkedIn banned the project's contributors. GodsScion's docs are upfront about it: "It is your responsibility to review and comply with them before using this tool with your account."

Tsenta doesn't apply through LinkedIn. it finds roles on company career pages and submits on the employer's own system, so there's no LinkedIn session in the loop for LinkedIn to flag.

tracking and tailoring

• AIHawk (2024): wrote success.json, failed.json and skipped.json to an output folder, and generated a résumé PDF per job from your YAML résumé.
• Auto_job_applier_linkedIn: CSV history of applied and failed jobs plus a local history page. it applies with the résumé you provide.
• neither script reads your inbox. recruiter replies land in your email and you match them to applications yourself.
• Tsenta: a tracker on every tier, including the free 25. recruiter replies are routed back to the right application and statuses move on their own. your résumé is tailored per role, with a cover letter where the employer accepts one, and with résumé review on, changes show as a diff you approve or edit. you choose how freely it rewrites (Honest or Aggressive), or turn tailoring off. the record of what was sent depends on the employer's system.

cost: free script vs paid agent

the scripts cost $0 to download. the real costs are elsewhere.

• LLM usage. AIHawk needed an LLM key for answers and résumés. GodsScion's AI is optional and off by default. hosted models bill per token: on OpenAI's pricing page today, gpt-4o-mini is $0.15 per million input tokens and $0.60 per million output tokens, and gpt-4o is $2.50 and $10.00, about 16 times more for the same work. Google's Gemini API has a free tier, but Google says free-tier content is used to improve its products. a local model through Ollama or LM Studio has no API bill but runs on your own hardware. neither project publishes tokens per application. AIHawk logged every LLM call to open_ai_calls.json, so you could check.
• your time. install, config, keeping the machine awake with Chrome open, and fixing things when LinkedIn changes its markup. GodsScion's own source code notes that the classes on LinkedIn's job cards are now "a rotating random string."
• your LinkedIn account. see the section above.

Tsenta's first 25 applications are free, no card. after that, Starter is $19 for 600 applications, Pro is $39 for 1,500, and Power is $99 for 4,500, per 30-day cycle. that's about 3 cents an application on Starter. only successful submissions count against your allowance, and quarterly or annual billing lowers the effective monthly rate by up to 36%. details on Tsenta pricing.

the comparison

	AIHawk (2024 code, via forks)	Auto_job_applier_linkedIn (GodsScion)	Tsenta
Status on Oct 30, 2026	LinkedIn code removed in 2025; repo archived Apr 16, 2026, now a different project	Active; last commit Sep 10, 2026	Managed service
License	Custom source-available (2024), AGPL-3.0 (2025), MIT for the new project	MIT (AGPL-3.0 before Aug 2026)	Commercial
GitHub stars	~32,300 on the renamed repo	~3,000	n/a
Where it applies	LinkedIn Easy Apply	LinkedIn Easy Apply; saves external links to a CSV	30+ employer ATSes, including Workday, Greenhouse, Lever, Ashby
Setup	Python, Chrome, virtual environment, three YAML files, LLM key	Python 3.10+, Chrome, launcher and local control panel	Sign up, upload résumé, set preferences
Runs on	Your computer	Your computer, Chrome in the foreground	Tsenta's cloud
Uses your LinkedIn account	Yes	Yes	No
Résumé tailoring	Generated PDF per job	Uses the résumé you provide	Per role, with optional diff review
Tracking	JSON files	CSV files and a local history page	Tracker with recruiter replies routed
Where you use it	Terminal	Browser control panel on your machine	Web, iOS, Android, iMessage, WhatsApp, Chrome extension, MCP server
Cost	Free plus LLM usage	Free; LLM optional	25 free; $19, $39, $99 per 30-day cycle
Who fixes it	Nobody upstream	One independent developer	Tsenta

who each one suits

pick an open-source script if:

• you're comfortable with Python, Chrome drivers and reading stack traces
• your target roles mostly post as LinkedIn Easy Apply
• you want everything on your own machine and want to read the code
• $0 matters more than your time, and you accept the risk to your LinkedIn account

if you go that way, start with GodsScion's, not an AIHawk fork. it's maintained, it has a control panel and a dry-run mode, it filters on sponsorship language, and the MIT license lets you change it however you like.

pick Tsenta if:

• the employers you want post on Workday, Greenhouse, Lever, Ashby or other ATSes, not only Easy Apply
• you don't want to keep a laptop awake with Chrome open
• you want a tailored résumé per role and a tracker that matches recruiter replies to applications
• you want to apply from your phone, iMessage or WhatsApp, or from Claude or ChatGPT through the Tsenta MCP server

where Tsenta is the worse pick: it isn't free past 25 applications, it isn't open source, it doesn't submit Easy Apply, and you're trusting a company with your profile. and Auto Apply, where the agent picks and submits matches without a per-role approval, is on Power only.

the closer

AIHawk proved people wanted an auto job applier on GitHub. it also showed the weak spot of building on LinkedIn: one platform can shut the whole thing down. in 2026, GodsScion's script is the honest open-source pick for people who want LinkedIn Easy Apply automation and accept the risk.

if your search runs through employers' own application systems, that's the part Tsenta was built for. try the 25 free applications, no card needed.

(we compare Tsenta with more tools in Tsenta vs competitors and the best auto job applier tools in 2026.)

questions people ask about AIHawk and Tsenta

What is AIHawk and how does it work?

AIHawk (Auto_Jobs_Applier_AIHawk) is an open-source Python project that, in its 2024 version, logged in to your LinkedIn account through Chrome, searched jobs using your config files, filled LinkedIn Easy Apply forms, and answered screening questions with an LLM such as OpenAI, Claude, Gemini or a local Ollama model. By March 2025 its README said the LinkedIn automation had been removed, and the GitHub link now redirects to a different project. Tsenta is a managed alternative that applies on employers' own systems.

Is AIHawk still on GitHub?

Not as a job applier. The original repository was archived on April 16, 2026. As of October 30, 2026, github.com/feder-cr/Auto_Jobs_Applier_AIHawk redirects to feder-cr/invisible_dots, a general self-hosted agent project whose commit history starts on September 29, 2026. The 2024 LinkedIn code survives only in forks such as 11844/Auto_Jobs_Applier_AIHawk, which have no upstream maintainer.

Is AIHawk free to use?

The code is free to download, but running it isn't free of cost. You need an LLM for answers and résumé generation, which means per-token API charges on a hosted model or your own hardware for a local one. You also spend time on setup and fixes, and you put your LinkedIn account at risk. The 2024 code shipped with a license that bars commercial use without the author's permission. Tsenta's first 25 applications are free with no card.

Does AIHawk work on Workday, Greenhouse, or Lever?

No. AIHawk's 2024 code automated LinkedIn Easy Apply only. GodsScion's Auto_job_applier_linkedIn also applies only through Easy Apply and saves external application links to a CSV for you to finish by hand. Tsenta submits on 30+ employer systems, including Workday, Greenhouse, Lever, Ashby and iCIMS.

Is using AIHawk against LinkedIn's terms of service?

LinkedIn's User Agreement prohibits using bots or other unauthorized automated methods to access its services, and its help center says members who use tools that automate activity on LinkedIn risk having their accounts restricted or shut down. AIHawk and Auto_job_applier_linkedIn both drive your own LinkedIn account in Chrome, so that risk applies. AIHawk's README said LinkedIn banned the project's contributors.

Can my LinkedIn account get restricted for using an auto job applier script?

Yes, that is the risk LinkedIn describes. LinkedIn also sets daily and speed limits on Easy Apply, pauses Easy Apply after rapid submissions, and says it may lower a member's daily limit when activity looks inauthentic. Tsenta doesn't apply through your LinkedIn account. It finds roles on company career pages and submits on the employer's own system.

How hard is it to set up AIHawk compared to Tsenta?

AIHawk's 2024 setup meant installing Python and Chrome, cloning the repo, creating a virtual environment, installing packages, and filling in three YAML files, including an LLM API key. GodsScion's tool is easier, with a double-click launcher and a local control panel, but it still runs on your computer with Chrome open. Tsenta needs an account, your résumé and your preferences, and the agent runs in Tsenta's cloud.

What is the best auto job applier in 2026?

It depends on where your target jobs live. For a free, local tool focused on LinkedIn Easy Apply, GodsScion's Auto_job_applier_linkedIn is a maintained open-source option, if you accept the account risk. For roles on employer systems like Workday and Greenhouse, Tsenta is the stronger choice because it finds roles, tailors your résumé, submits and tracks replies across 30+ ATSes. More options are in the best auto job applier tools in 2026.

Does Tsenta apply through LinkedIn?

No. Tsenta finds roles on 50,000+ company career pages and submits on the employer's own application system, such as Workday or Greenhouse. It doesn't apply through your LinkedIn account or submit LinkedIn Easy Apply forms, so a role that exists only as an Easy Apply posting isn't one Tsenta can send.

How much does Tsenta cost compared to running AIHawk?

An AIHawk fork costs nothing to download, plus LLM usage, your time and the risk to your LinkedIn account. Tsenta's first 25 applications are free with no card. After that, Starter is $19 for 600 applications, Pro is $39 for 1,500 and Power is $99 for 4,500 per 30-day cycle, about 3 cents an application on Starter. Only successful submissions count.

Who should use an open-source job applier instead of Tsenta?

Developers who are comfortable with Python, want to read and change the code, want their data on their own machine, and mostly target LinkedIn Easy Apply roles. If that's you and you accept the account risk, GodsScion's Auto_job_applier_linkedIn is a better starting point than an unmaintained AIHawk fork. If your target employers post on Workday, Greenhouse or Lever, Tsenta covers what the scripts don't.

Which auto job applier is best for international students on OPT?

They help in different ways. GodsScion's script can skip LinkedIn postings whose text says they don't sponsor visas. Tsenta surfaces the sponsorship signals it knows about and uses your work authorization when answering application questions, across employer systems. Neither can guarantee sponsorship, because that is the employer's decision. More on Tsenta for international students.