Skip to content

A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source and free forever. Try it out today!

20.3 KB
Raw Blame
---
title: "Adding an application"
description: "Save a job from a link, pasted posting or optional keyword search, then prepare a resume or record what you already submitted."
---

Save a job as soon as you find one you like, or record an application you've already submitted. Paste the job link or posting text, check the role and company, and save it. A saved posting gives [Check](/guides/checking-your-resume), [the assistant](/guides/using-the-assistant) and [your cover letter](/guides/writing-a-cover-letter) context for this job.

Saving a job and preparing a resume manually work without an AI connection or an external web account.

## Add an application

<Steps>
  <Step title="Open the Save a job dialog">
    In **Applications**, select **Save job**. You can also press <kbd>⌘</kbd> <kbd>K</kbd> (<kbd>Ctrl</kbd> <kbd>K</kbd>{" "}
    on Windows and Linux) and run **Save job**.
  </Step>
  <Step title="Paste the job link or posting">
    Paste into **Job link or posting text**. A single web address counts as a link; anything else counts as posting
    text. Select **Read posting** to retrieve a link's text and any available job details. Reading starts only when you
    select the button; you can also save the link without reading it. If you close the dialog before saving, what you
    pasted and what was read are still there when you open it again in the same tab.

    Before reading a link, the guidance says **Read posting to fill the role and company.** This is an invitation to
    read, not a result. An empty form shows manual-entry guidance instead. After reading, review the details found or
    follow the warning if the page was empty or could not be read.

  </Step>
  <Step title="Check the role and company">
    Reading may fill **Role**, **Company**, **Location** and **Salary**. Correct anything that was found, and fill
    missing fields yourself. Only **Role** and **Company** are required. When AI read the posting, **Show what it asks
    for** lists the requirements it found.
  </Step>
  <Step title="Pick a stage">
    **Saved** is the default. Choose **Already applied** or **Already interviewing** only if that happened, then confirm
    the application date and select any documents you actually submitted. Your own Saved, Applied and Interview stages
    appear here too. If you applied to the same company before and never heard back, those applications are listed as a
    reminder; you can still save.
  </Step>
  <Step title="Add it">
    Select **Save job** to open its details, or **Record application** when recording a submission. For a saved job,
    **Resume** starts on your most recently edited resume that isn't a copy for another job. With a resume chosen,
    **Save and prepare resume** makes a copy for this job, named from the resume and the company, and opens it; your
    original stays unchanged. With **No resume selected**, it opens the copy step instead. See [Tailoring a resume for a
    job](/guides/tailoring-a-resume-for-a-job).
  </Step>
</Steps>

On Android and iPhone, **Applications → Save job** opens the same preparation choice after you review the posting. Under **More details**, **Resume** lets you change the source. **Save and prepare resume** copies the selected resume for the saved job and opens **Write**; with a usable AI connection, it also opens the assistant's preparation suggestions. Nothing is generated until you choose a suggestion. Without a resume, the existing copy step lets you import or create one. If copying fails after the job is saved, that step opens for the same saved job so you can continue. Leaving an unsaved native form still asks before discarding it. When adding a job inside another document picker, **Save job** returns the job to that picker.

<Frame caption="Web reference: a fictional posting and source resume, ready for Save and prepare resume. Android and iPhone reuse the same copy service; this image shows the web dialog.">
  <img
    src="/images/guides/adding-an-application/save-prepare-review.png"
    alt="Save a job dialog with posting text, Role and Company fields, stage choices, Location, Salary, a resume selector, and Save job and Save and prepare resume buttons"
  />
</Frame>

<Frame caption="An unread link offers Read posting; it does not claim the page has already been checked.">
  <img
    src="/images/guides/adding-an-application/save-job-unread.png"
    alt="Save a job dialog with an unread fictional posting link, Read posting button, empty role and company fields, and guidance to read the posting to fill them"
  />
</Frame>

<Frame caption="The same unread-link guidance at phone width in the web app.">
  <img
    src="/images/guides/adding-an-application/save-job-unread-phone.png"
    alt="Phone-width web Save a job dialog with Read posting guidance before a link has been read"
  />
</Frame>

## Capture from your browser

In **Save a job**, select **Download browser helper**. Unzip the download, open `chrome://extensions` (or `edge://extensions`), enable **Developer mode**, choose **Load unpacked**, and select the unzipped folder. This helper supports current desktop Chrome and Edge. Other browsers, restricted pages, and embedded frames can use URL or pasted-text import.

Open the helper and **Connect** your installation origin, such as `https://rxresu.me` or `https://resume.example.org`. It stores only that origin. Sign in inside the installation itself; never paste API keys, browser cookies or passwords into the helper. Self-hosted installations need HTTPS; HTTP localhost is allowed for development.

On one posting, open the helper and choose **Capture this posting**. A page's single structured JobPosting supplies role, company, location and description without AI. If unavailable, select only the posting text and capture again. The helper never falls back to sending the whole page. Selected text excludes hidden content and editable notes. Expand any closed job details before selecting their text; a closed disclosure contributes only its visible summary. The helper removes known tracking and session parameters while preserving supported job identifiers. If unknown query parameters could change the job identity or carry credentials, capture stops and asks you to enter the URL or text explicitly. Different jobs with the same title and company are not treated as one structured posting.

Review and correct the URL, **Role**, **Company**, **Location** and **Saved description** in the installation, then choose **Save job**. When already tracked, use **Open tracked job** or **Update posting details**. Updating this way preserves stage, notes and submitted documents. An absent location or description keeps the existing value; use Edit details to clear it intentionally. Capture does not contact the employer, submit an application or mark it applied.

Choose **Disconnect** in the helper to remove its installation setting. Reconnect explicitly to change installations. No posting history or credentials are retained in the helper. Capture opens the review through a URL fragment containing the posting text. The app removes that fragment from its address bar, but the browser may retain it in local or synced browsing history. Use the helper only for text you are comfortable retaining there. Closing or reloading the review before saving discards the capture; sign in first if your sign-in provider needs a full page redirect, then capture again.

<Frame caption="A captured fictional Lumen Workshop posting, corrected before saving">
  <img
    src="/images/guides/career/posting-capture.png"
    alt="Save a job review showing the fictional Lumen Workshop role, editable captured description and location, before saving"
  />
</Frame>

## Import a mobile share

In the current native iOS or Android app, connect to your installation using the existing **Another server** sign-in flow. Use the browser's **Share** action and select **Reactive Resume**. Share one URL or the posting text you deliberately selected. On iOS, choose **Keep for review, then reopen app**, return to your Home Screen, and reopen Reactive Resume yourself within 15 minutes. The Share extension finishes normally; it does not launch the app. Open the app and sign in before sharing; after a day without opening it, open it again before sharing. On Android, selecting Reactive Resume opens the review directly. Review the destination installation and imported text; use **Read posting** for a public URL or fill missing fields manually. Save only after correcting the details. Duplicate controls work as above. If the duplicate check is unavailable on an older self-hosted server, saving remains available. Check Applications manually before saving or upgrade the server to enable duplicate detection.

Mobile browsers choose what their share sheet exposes. Many send only a URL, without the page title, description or signed-in page contents. The app receives only the shared payload; it cannot read the browser's session. For a session-only posting, select and share its text, or paste it into **Applications → Save job**. Files and multiple postings are unsupported. A new store build is required for the native receive target; older binaries retain URL/paste import.

Switching server or signing out disconnects the native app through its existing account controls. Captured drafts are account scoped and cleared on identity changes. On iOS, pending raw text stays in the private app group until one-time review, cancellation or account/server change. After 15 minutes it is no longer available for review and is removed on the next app activation. Closing review discards it. A clipped description shows a warning and retains incomplete-source metadata when saved. No capture creates a separate server record: saved posting fields use the existing application, account export and account deletion flows.

## Find a posting with keyword search

If search is available on your installation, the **Save a job** dialog also shows **Search job postings**. Search needs a configured **Web access** connection or server-provided search; an AI provider alone does not enable it. Check **Settings → AI & developer → Web access**, or ask your administrator if the server provides search. You can always paste a link or posting instead.

<Steps>
  <Step title="Search for a role">
    Enter a role, company or location under **Search job postings**, then select **Search**. Search runs only when you
    select the button or press Enter in that field.
  </Step>
  <Step title="Choose a posting">
    Review the result title and description. **View posting** opens the source page in a new tab. Select **Use posting**
    to put its address into **Job link or posting text**. This clears any role, company, location and salary already in
    the dialog; it does not read or save the posting.
  </Step>
  <Step title="Read and review">
    Select **Read posting**, then review **Saved description** and the role, company, location and salary. Correct
    missing or inaccurate details. If the page cannot be read, use **Paste description instead**. The reason stays above
    the paste field: a board's list of jobs says to open one posting, and other failures name what to do next. Opening
    the paste field never pretends the read succeeded.
  </Step>
  <Step title="Save the job">
    Choose the appropriate stage, then select **Save job**, **Record application** or **Save and prepare resume**, as
    described above. Search results are not saved as applications until you choose to save.
  </Step>
</Steps>

Search sends your search terms to the configured web search service. Reading a link retrieves its public page through the selected web reader or the built-in reader. **Read posting** reads structured page details without AI. **Read with AI** is optional and names the connection receiving the posting before you choose it. Saving a job never submits an application or contacts the employer.

If search fails or finds nothing useful, change your keywords or paste a posting link directly. For searches that run on a schedule, see [Finding roles automatically](/guides/finding-roles-automatically).

Search failures name a rejected key, an exhausted quota, a temporary request limit or an unavailable service. A personal connection offers **Set up web access** in place: repair it without leaving your query, posting or edited fields. Rejected-key and quota failures remain visible after reloading. For a request limit, wait before trying again. If your server provides search, ask its administrator to repair the service; personal key controls cannot change it.

<Frame caption="Web search recovery keeps the fictional query and offers personal connection repair in place.">
  <img
    src="/images/guides/career/search-key-recovery.png"
    alt="Web Save a job dialog with a rejected Firecrawl key warning, Set up web access button and the fictional search query retained"
  />
</Frame>

After a rejected key or exhausted quota, **Retry with _connection name_** appears when your default tested AI connection supports native search and your server supports the fallback. Direct OpenAI, Anthropic and Google Gemini connections are supported; custom endpoints and compatible gateways are excluded, and individual models can still refuse. Review the named connection, model, query-only data scope and possible charges, then approve this operation. Declining keeps your query and recovery controls. A successful retry returns retrieved source links without marking the failed web connection ready.

<Frame caption="An explicitly approved retry returns a fictional posting while preserving the authored job fields. This is the web interface.">
  <img
    src="/images/guides/career/search-native-retry.png"
    alt="Save a job shows Retry with Fictional search connection, a retrieved Fictional design engineer result, and retained Accessibility engineer and Fictional studio fields"
  />
</Frame>

Pasted postings and public-link reading remain available while search needs attention. Public-link reading can fall back to the built-in reader. Keyword search uses the configured search service; it does not silently switch to an AI provider's search.

## What gets read from a posting

After a successful link read, **Saved description** shows a relative read time and the host retained by the reader,
such as **Read 3 minutes ago from jobs.example.test**. The time updates while the review is open. It says when the app
read the page, not when the employer last updated it; the employer may have changed or closed the posting since.
**View posting** keeps that receipt after saving, even if you later edit the application's link. Older saved postings
without a retained host show the generic posting-page label; missing or invalid read times never become a claimed read.
Pasted descriptions remain labeled **Pasted description**.

<Frame caption="A fictional posting read, reviewed before saving. The receipt names the retained source host and elapsed time.">
  <img
    src="/images/guides/career/posting-read-review.png"
    alt="Save a job review for Lumen Workshop showing Read 3 minutes ago from jobs.example.test and a warning that the employer may have changed the posting"
  />
</Frame>

<Frame caption="The retained read receipt stays with the saved description. Editing the job's link does not relabel its source.">
  <img
    src="/images/guides/career/posting-read-saved.png"
    alt="Saved fictional Lumen Workshop posting showing its edited link and the original jobs.example.test read receipt"
  />
</Frame>

**Read posting** uses structured page details without AI. The app first confirms that your server supports reading without AI. If that check fails, it sends no parse request: retry, update your server, or enter the details manually. Optional **Read with AI** remains a separate choice and uses your selected [AI connection](/guides/using-ai), disclosed beside the button:

| You paste    | Without an AI provider                                                                | With optional Read with AI                                      |
| ------------ | ------------------------------------------------------------------------------------- | --------------------------------------------------------------- |
| A job link   | Saves the page text; may fill role, company and location from structured job details. | May also fill salary and a short list of what the job asks for. |
| Posting text | Saves the text. Fill in the role and company yourself.                                | May fill role, company, location, salary and requirements.      |

Pasted text is saved up to 20,000 characters. A link keeps its original address; its description stays empty until you read it or paste text. A long **Saved description** scrolls inside its own box, so the role, company and save buttons stay reachable.

<Frame caption="Long pasted text and the saved description each scroll inside their own box. Role, company and Save job stay reachable below them.">
  <img
    src="/images/guides/adding-an-application/long-description.png"
    alt="Save a job dialog with a long posting in a scrollable text box, a clipped-description warning and the start of the separate Saved description field"
  />
</Frame>
Review **Saved description** after reading. If the description is clipped or incomplete, use **Paste description
instead** to replace it before preparing your resume; the original link is retained. Source details such as retrieval
time appear when available, and do not guarantee the posting is current.

When a page publishes structured job data, its own title and company are kept over anything a model suggests, and escaped line breaks in its description are decoded. The employer's exact title is what their application system and emails use. Application forms, their buttons and selectors, and equal-opportunity or veteran boilerplate are left out of the saved description; the role's own text is kept.

AI enrichment is optional. If it fails, the retrieved description is retained so you can complete the fields manually. When you pasted the posting text yourself and AI couldn't read it, the text stays in **Job link or posting text** and is saved from there; no second **Saved description** box appears. One status explains the failure. In the web dialog, the save actions stay at the bottom while you scroll the fields.

<Frame caption="An unsuccessful optional AI read keeps one warning and the original posting. The web save actions stay visible while the fields scroll.">
  <img
    src="/images/guides/adding-an-application/posting-ai-recovery.png"
    alt="Scrolled web Save a job dialog with the retained fictional posting, one AI failure warning, manually completed Tools Engineer and Fictional Studio fields, and visible Save job and Save and prepare resume buttons"
  />
</Frame>

<Note>
  Only public HTTP(S) pages can be read. If a page requires sign-in or can't be retrieved, paste its description in
  **Paste description instead**, fill in the role and company, and save. The original link stays with the job.
</Note>

## Add more details

The Add dialog keeps things short on purpose. Once the application is saved, its details sheet opens, where you can add the salary, source, contacts, tags, notes and a follow-up date. To edit everything in one form, including the job posting link, location and linked resume, open the **⋯** menu in the details sheet and select **Edit details…**. See [Managing an application](/guides/managing-an-application).

## Other ways to add applications

- **From a spreadsheet**: import many at once from a CSV file in the web or native mobile app. Check the column matches and summary before importing. See [Importing applications from CSV](/guides/importing-applications-from-csv).
- **From an AI client**: create applications over MCP. See [Managing applications with MCP](/guides/managing-applications-with-mcp).

## Next steps

- [Managing an application](/guides/managing-an-application): move it through stages and keep notes.
- [Tailoring a resume for a job](/guides/tailoring-a-resume-for-a-job): make a version of your resume for this posting.

## Re-read a posting

Reading a different posting replaces values filled by the previous read. Fields you typed or deliberately cleared stay yours, including edits made while the read runs. **Read posting** asks you to approve the named web reader and uses no AI. Optional **Read with AI** names the AI connection receiving the posting. Without an AI provider, public links can still use the built-in reader, or you can save the role and company manually. **Set up AI** keeps your input in place. A rejected web key shows **Needs attention** and repair controls; pasting the posting stays available.