Important
I'm looking for a job right now! If you know of any openings that match my skill-set, please let me know! You can read my resume over at my cv project. Thanks!!
smart-letters is a CLI application designed to help you write cover letters rapidly
for specific job postings. By harnessing the power of OpenAI and Python, we can generate
a styled cover letter in PDF format in seconds. The slowest part is waiting for the
request from OpenAI to return!
pip install smart-lettersTo see all the configuration options, run:
smart-letters config --helpThe simplest working config would look like this:
smart-letters config bind --openai-api-key=<your-api-key> --resume-path=<path-to-resume-text> --candidate-name=<your-name>To see all the generation options, run:
smart-letters generate --helpAn example run command would look like this:
smart-letters generate --company=ACME --position="Senior Engineering Lead" https://github.com/dusktreader/smart-letters/blob/main/etc/fake-listing.mdFollow the prompts and get your ready-to-send cover letter!
There's a lot of pain involved with job searching these days. There's the struggle to find postings that match your experience, expertise, and preferences. Then, there's the frustration of all the different web-apps that are used to gather applications. Of course, there's the rejection which is often automated; that is, if you get any rejection notice instead of just silence.
However, probably the most frustrating part is the composition of cover letters. In the current marketplace for tech talent, you need to be blasting out lots and lots of applications. I've read horror stories on Reddit of seasoned devs that have sent out literal thousands of applications over a several month span as they searched for their next opportunity. Obviously, this is a process that needs to be fine tuned to maximize production and maintain quality. The cover letter is a wrench thrown directly into the machinery of the application process.
I understand that hiring teams may feel that cover letters are a useful tool. It helps to screen out endless spam from automation. It's a way to ensure that the candidate has at least read the posting. Plus, it gives you a snapshot into the personality and motivation of the person applying. Sometimes I worry that the main reason that cover letters are requested is that it takes only ticking a single check-box in the submission form to many of these application SaaS platforms. In any case, writing the cover letter is the main hot-spot of the application process.
Obviously, AI is a powerful tool for producing a lot of text very quickly. But, cover letters produced by AI...read like they were written by AI. If you really want to get good results, you need to tailor the letter to the job posting, refine your prompt, feed it with personal information, refine the output, and then edit it by hand afterward to get a good result.
Even that is quite boring and time consuming. So, as most good Python engineers like to do, I decided to automate the boring stuff!
The smart-letters program stores its configuration in a file so that it can use
the same settings for many runs without having a super-cluttered command line. You can
check out the location where the config file is saved in the config.py module.
There are several sub-commands that you can use to interact with your configuration:
This is the core command. We used it in the Quickstart. This sub-command binds the
provided configuration options to the smart-letters program by storing them in a
config file.
Sometimes you want to just change one or two configuration options instead of supplying
all of them again. For that, you would use the update subcommand:
smart-letters config update --filename-prefix=awesome-letterIn this case, we only updated the filename prefix setting and left all the others as they are.
If you need to just unset one of the config options, you can do that with the unset
subcommand. You don't need to provide any value, just pass the option:
smart-letters config unset --filename-prefixIf you want to see the current configuration for smart-letters, use the show command:
smart-letters config showIt will produce some nicely formatted output that would look like this:
╭──────────────────────────── Current Configuration ───────────────────────────────────╮
│ │
│ openai_api_key -> <omitted> │
│ resume_path -> /home/coyote/git-repos/personal/cv/README.md │
│ candidate_name -> Wile E. Coyote │
│ filename_prefix -> cover-letter │
│ heading_path -> /home/coyote/text/heading.md │
│ sig_path -> /home/coyote/images/sig.png │
│ │
╰──────────────────────────────────────────────────────────────────────────────────────╯
If you want to just completely clear out your config, use the clear subcommand:
smart-letters config clearThere are some notable settings that weren't used in the Quickstart. Let's go over them here!
This one is really simple. By default, smart-letters will use the generic
"cover-letter" prefix for your file names. If you want to change it to something else,
you can set that here.
I, personally, like to include a heading in my letters that includes my name and some basic contact information. You might want to supply one as well. You should note that the contents of the letter are formatted as Markdown before they are rendered to a PDF, so if your heading is formatted as markdown, it will get rendered nicely. Here's an example heading:
# Wile E. Coyote
📍 [Tucson, AZ](https://maps.app.goo.gl/5siCgp4pUymGPU499) /
📧 [wile.e.coyote@gmail.com](wile.e.coyote@gmail.com) /
🛠 [wile.e.coyote@github](https://github.com/wile.e.coyote)
---I also like to include a signature image in my generated cover letter. If you include one with this option, it will be placed in between your closing and your name:
Best regards,
[SIG HERE]
Wile E. Coyote
This option lets you set a target directory where letters will be saved when they are generated.
If you don't supply it, the default will be the current directory where you call the smart-letters
app.
By default, smart-letters uses a built-in developer prompt when it's generating letters. If you
want to have more control of the prompt that is used, you can supply a path to your own prompt
file.
By default, smart-letters uses a built-in user prompt that is a mako template. The template has
the following values rendered into it at generation time:
- resume_text: The raw text of your resume
- posting_text: The raw text of the job posting
- example_text: An optional example letter to use as reference
You can supply a path to your own template if you want to have more control of the user prompt.
This option allows you to tell smart-letters which app to use to edit the generated letters
before they are finalized. If you don't supply it, smart-letters will attempt to use the
default editor configured for your system.
There are a few options that are available on the generate sub-command that can be
useful. Let's go over them quickly:
This is simply the name of the company for which the letter will be generated. It is used in creating the salutation and setting the filename.
If it's omitted, the salutation will just be generic:
Dear Hiring Manager,
If you set it, however, the salutation will incorporate it:
To the Hiring Team at <company>:
This option is only used for inclusion in the filename. It's useful to be able to distinguish between many letters that you've generated over time.
This option can be used to give the generate command an example (text) letter to
use for a reference when it's generating your cover letter.
By default smart-letters will ask you if you want to render the final letter to PDF
when it's done building it. With this option, you can control the answer to this
question ahead of time.
When you are debugging smart-letters, the most time-consuming part is waiting for
OpenAI to generate the letter. If you want to skip that part, you can just pass this
option, and it will use a pre-baked letter body instead of calling out to OpenAI.
The smart-letters program stores its configuration in a file so that it can use
the same settings for many runs without having a super-cluttered command line. You can
check out the location where the config file is saved in the config.py module.
If you already have a Markdown letter that you would like to render to PDF, you can
use the render command. It will produce the letter by converting the Markdown to
HTML, styling it with CSS, and then rendering the result to a PDF.
There is options available on the render sub-command:
By default, render will use the same file stem as the provided Markdown file. However,
if you want to change the stem, you can use this option.
For example, if the provided markdown file is named my-letter.md, the rendered PDF will
be named my-letter.pdf.