Getting your resume aka an CV (ANSI-v 🤡) straight to your and anyone else's terminals:
Be warned though, for this is kinda useless and just for fun:
-
Create your resume according to the JSON Resume Schema (see also the schema specification) either:
- use ChatGPT (or another LLM) with the following prompt (you need to fill in the spaces for
[resume]
and[json_schema]
):Note: forResume:[resume] JSON Resume Schema:[json_schema] Provide JSON data structure of the resume, formatted according to the JSON Resume Schema Output json, no yapping
json_schema
you can just use the example from here - manually (see the
heyho
sample for a possible starting point), - exporting from LinkedIn using Joshua Tzucker's LinkedIn exporter (repo)1, or
- exporting from one of the platforms advertised as offering JSON resume integration.
- use ChatGPT (or another LLM) with the following prompt (you need to fill in the spaces for
-
Create a public gist named
resume.json
with your resume contents. -
You're now the proud owner of an ancv. Time to try it out.
The following examples work out-of-the-box. Replace
heyho
with your GitHub username once you're all set up.-
curl:
curl -L ancv.io/heyho
with
-L
being shorthand for--location
, allowing you to follow the redirect fromhttps://ancv.io
through tohttps://ancv.io
. It's shorter than its also perfectly viable alternative:curl https://ancv.io/heyho
Lastly, you might want to page the output for easiest reading, top-to-bottom:
curl -sL ancv.io/heyho | less
If that garbles the rendered output, try
less -r
aka--raw-control-chars
. -
wget:
wget -O - --quiet ancv.io/heyho
where
-O
is short for--output-document
, used here to redirect to stdout. -
PowerShell 7:
(iwr ancv.io/heyho).Content
where
iwr
is an alias forInvoke-Webrequest
, returning an object whoseContent
we access. -
PowerShell 5:
(iwr -UseBasicParsing ancv.io/heyho).Content
where
-UseBasicParsing
is only required if you haven't set up Internet Explorer yet (yes, really). If you have, then it works as PowerShell 7 (where that flag is deprecated and the default anyway).
-
All configuration is optional.
The CV is constructed as follows:
In summary:
-
you control:
-
the template.
Essentially the order of items, indentations, text alignment, position of dates and more. Templates are like layouts/skeletons.
-
the theme.
This controls colors, italics, boldface, underlining, blinking (yes, really) and more. A couple themes exist but you can easily add your own one.
-
the language to use.
Pre-set strings like section titles (Education, ...), names of months etc. are governed by translations, of which there are a couple available already. All other text is free-form.
-
text content like emojis and newlines to control paragraph breaks.
Emojis are user-controlled: if you want them, use them in your
resume.json
; in the future, there might be templates with emojis baked in, but you'd have to actively opt into using one. -
date formatting, in a limited fashion through a special
dec31_as_year
toggle. If that toggle istrue
, dates in the formatYYYY-12-31
will be displayed asYYYY
only. -
lastly, there's a toggle for ASCII-only output.
It only concerns the template and controls the drawing of boxes and such (e.g.,
-
versus─
: only the latter will produce gapless rules). If you yourself use non-ASCII characters in your texts, use a language containing non-ASCII characters (Spanish, French, ...) or a theme with non-ASCII characters (e.g., a theme might use the•
character to print bullet points), non-ASCII Unicode will still occur. As such, this toggle currently isn't very powerful, but with some care it does ultimately allow you to be ASCII-only.
If you come up with new templates, themes or translations, a PR would be highly appreciated.
-
-
you do not control:
-
anything about a viewer's terminal!
Any recent terminal will support a baseline of features (e.g., colors), but large parts of the functionalities depend on the font used: proper Unicode support is needed for pretty output (see
ascii_only
), and ideally emojis if you're into that (although it's easy to pick an emoji-free template). Many themes leverage Unicode characters as well. -
access to your CV: like the gist itself, it will be publicly available on GitHub.
-
Configuring ancv
requires going beyond the vanilla JSON Resume schema.
You will need to add an (entirely optional) $.meta.ancv
field to your resume.json
.
The provided schema will be of help here:
an editor capable of providing auto-completion based on it, like Visual Studio Code, will make filling out the additional configuration a breeze.
The schema will further inform you of the default values (used for unspecified fields).
Since everything is optional, a valid JSON resume (without an ancv
section) is valid for ancv
use as well.
Install the package as usual:
pip install ancv
This also allows you to import whatever you could want or need from the package, if anything. Note that it's pretty heavy on the dependencies.
See also the available packages aka images:
docker pull ghcr.io/alexpovel/ancv
Versioned tags (so you can pin a major) are available.
Once installed, you could for example check whether your resume.json
is valid at all (validate
) or get a glimpse at the final product (render
):
# pip route:
$ ancv render resume.json
# container route:
$ docker run -v $(pwd)/resume.json:/app/resume.json ghcr.io/alexpovel/ancv render
Self-hosting is a first-class citizen here.
The https://ancv.io site is hosted on Google Cloud Run (serverless) and deployed there automatically, such that the latest release you see here is also the code executing in that cloud environment.
That's convenient to get started: simply create a resume.json
gist and you're good to go within minutes.
It can also be used for debugging and playing around; it's a playground of sorts.
You're invited to use this service for as much and as long as you'd like. However, obviously, as an individual I cannot guarantee its availability in perpetuity. You might also feel uncomfortable uploading your CV onto GitHub, since it has to be public for this whole exercise to work. Lastly, you might also be suspicious of me inserting funny business into your CV before serving it out. If this is you, self-hosting is for you.
For simplicity, using Docker Compose (with Docker's recent Compose CLI plugin):
- Clone this repository onto your server (or fork it, make your edits and clone that)
cd self-hosting
- Edit Caddy's config file (more info) to contain your own domain name
- Place your
resume.json
into the directory - Run
docker compose up
Caddy (chosen here for simplicity) will handle HTTPS automatically for you, but will of course require domain names to be set up correctly to answer ACME challenges.
Handling DNS is up to you; for dynamic DNS, I can recommend qmcgaw/ddns-updater
.
If you self-host in the cloud, the server infrastructure might be taken care of for you by your provider already (as is the case for Google Cloud Run). In these cases, a dedicated proxy is unnecessary and a single Dockerfile might suffice (adjusted to your needs). True serverless is also a possibility and an excellent fit here. For example, one could use Digital Ocean's Functions. If you go that route and succeed, please let me know! (I had given up with how depressingly hard dependency management was, as opposed to tried-and-tested container images.)
Footnotes
-
The exporter has a couple caveats. You will probably not be able to paste its result into a gist and have it work out of the box. It is recommended to paste the export into an editor capable of helping you find errors against the contained
$schema
, like VS Code. Alternatively, a localancv render your-file.json
will printpydantic
validation errors, which might be helpful in debugging. For example, the exporter might leave$.basics.url
an empty string, which isn't a valid URI and therefore fails the schema and, by extension,ancv
. Similarly,endDate
keys might get empty string values. Remove these entries entirely to stay conformant to the JSON Resume Schema (to whichancv
stays conformant). ↩