What's new?

The directory could only ever tell you whether a tool was reachable by a machine at all: it published an llms.txt or an MCP server, or it did not. That is a useful answer, but it treats every tool on the list as the same. Tools now carry a letter grade instead, so you can see at a glance how far a product has actually gone.

What's new?

• Tool pages show an agent-readiness grade beside the repository, docs, and MCP links.
• The full scoring is published on the agent-ready page: every signal, what it is worth, and where each band starts.
• Grades are worked out from checks we already run against each product, and recalculated every time a page loads, so a grade can never describe a product as it used to be.
• Nothing is self-reported. The only way to change a grade is to change the product.
• Tools that publish an MCP server you run locally now get credit for it, rather than being ranked below tools that happen to host theirs.

We only publish a grade of C and above. Not publishing an llms.txt or an MCP server is a perfectly ordinary thing for a good tool to be, and this directory is not here to put a failing mark on someone else's work.

Products get acquired, newsletters go quiet, and communities get absorbed into bigger ones. A directory that does not say so is quietly wrong. This is the first pass of a recurring check, and it introduces a new block on listing pages: a short, dated opinion from us about where something actually stands.

What's new?

• Stainless has been delisted. It was acquired in May 2026 and its hosted SDK generator is winding down, so new signups are closed and there is nothing left to adopt.
• Developer Avocados Weekly now says plainly that its last issue was May 2024, and that the value is the archive rather than a subscription.
• Developer Marketing Alliance now says that its site redirects to Product Marketing Alliance, so you know where you will land before you click.
• Every note carries the date we wrote it, so you can judge how current the opinion is.
• Hashnode's entry now records its pricing model, checked against its own pricing page.

We only delist when there is genuinely nothing left to use. A dormant newsletter with a good archive stays, clearly labelled.

There is a lot of good DevRel talk material online and very little of it is structured. Archives exist, and deeper ones than this, but they are card walls: you cannot ask them how long a talk runs or which edition it came from. DevRel talks is a first pass at the version you can actually query.

What's new?

• 73 talks from DevRelCon London 2023 and 2019, each one linking to the video on the conference's own channel
• Every row was checked twice against the conference's published playlist, and titles are stored exactly as the publisher wrote them, typos included
• Runtimes are shown, so you can tell a keynote from a lightning talk before you click
• All 73 talks are in the machine-readable version of the site too

We index talks, not speakers

This is a record of works, and we have built it so it cannot drift into being a database of people. There is no speaker field, no speaker page, and no way to filter by person. A conference publishes a talk under a title that usually names its speaker, so attribution stays in the title exactly as the conference wrote it, and nothing here turns that into a separate record of an individual.

Our automated checks fail the build if anyone adds a person-based field or page, so this is enforced rather than just promised. How we curate sets out the full policy.

If you spoke at one of these talks and would rather not be indexed here, mail the address on our privacy page and it comes out within three business days, usually the same day. No reason needed.

"Do they cover travel?" is the question that decides whether a working speaker submits a talk, and almost nothing on the internet answers it in a form you can scan. We read all 21 currently-open CFP pages on the events page and wrote down what each one actually says.

What's new?

• Chips on the event rows where a CFP offers something: travel covered or part-covered, accommodation, an honorarium.
• A "what the CFP says speakers get" block on every event page with a CFP, including the cases where the answer is no.
• The date each page was read, next to the claim, because organizers change these terms without notice.

The honest numbers

Of the 21 open CFPs, 8 say something about speaker costs and 13 say nothing at all.

That is not us giving up on the other thirteen. We read every one of them, and the page genuinely does not mention travel, accommodation, or a fee. The event pages say so explicitly, because "we checked and the CFP is silent" is a different and more useful answer than a blank space.

Silence is the norm here, not a red flag. Plenty of organizers help speakers who ask, and simply never wrote it on the page.

What these labels do not mean

They describe what the page says, nothing more.

A CFP that does not mention travel has not refused to help - it just did not say, and asking is usually worth it. Where a page does state a no, the event page phrases it as "not covered, per the CFP page" rather than as a verdict on the organizer, and that phrasing never appears as a chip in the list.

Nothing here is negotiated on your behalf, and nothing is estimated. Several events offer capped or conditional support - a travel refund up to a limit, a room for two nights, help "case by case" - and those show as part-covered rather than covered, because rounding a cap up to a promise is how someone ends up out of pocket. Confirm with the organizer before booking anything.

A job posting tells you about one role. It cannot tell you whether that company hires for DevRel regularly or posted once and went quiet, and their careers page cannot tell you either, because it only ever shows today. We have been reading these job boards twice a week and keeping a record, so now the record has somewhere to live.

What's new?

• Who hires for DevRel: every employer whose board we track, with their open roles
• Each company page shows what is open now and the date we started watching their board
• Once a posting comes down, it stays on the company page with the date it left, which is the part no live job board keeps
• Job pages link through to the employer, so you can go from one role to everything we have seen from that company
• Where a company also has a directory entry, the two are linked together

We only record what a posting said and when we saw it, never anything about who applied. Company pages with a single posting and nothing else are kept out of search results, because a page that thin is just the job posting again.

The most-quoted pay figure in Developer Relations sits behind a form, dates from 2023, and never says what it leaves out. There is now a free alternative that does: DevRel salary data, rebuilt from job postings every day.

What's new?

• The headline figures, computed over every open role on the board that states pay in USD: median asking range, lowest and highest stated bound, and the share of roles that disclose anything at all.
• All 26 individual ranges, published. Not just the aggregate. Every disclosing role appears with the band its posting stated, drawn on a shared scale so you can see the whole distribution instead of trusting one number.
• A remote, hybrid, and on-site cut, with the stated bounds for each arrangement.
• A published methodology that names the five things you need in order to judge the figures, including the ones that make them look worse.

What the numbers are not

These are asking ranges, not pay. Every figure comes from a band an employer published in a job posting: no accepted offers, no equity, no bonuses, and nothing self-reported. The offer that follows a posting often is not the midpoint of its range.

Nothing is estimated. A range appears only if the posting stated it, and roles quoted in other currencies are left out rather than converted. Where the sample is too thin to support a figure, the figure is withheld instead of published with a caveat - which is why some cells read "n too low" rather than showing a number we would not stand behind.

The sample is also a watchlist of companies known to hire in DevRel, not a census of the field, so it skews toward US-headquartered, well-funded employers. Every number on the page states how many roles it was computed from, so you can judge it yourself.

Also in this release

• The salary figures shown on the job board and in the numbers now link to that methodology, so an aggregate never travels without a route to how it was made.
• The same data is available as Markdown at /salaries.md for anyone building on it, and through the site's MCP server.

The directory has always listed tools one at a time, which is useful once you know what you are looking for and not much help before that. Stacks are the other half: for a given job, the decisions in the order you actually have to make them, with the shortlist we have verified for each one.

What's new?

• Three stacks to start: documentation, community, and measurement
• Each step explains the decision first and then names the candidates, including which option suits which situation
• Every stack links to the guide behind it, the open roles doing that work, and upcoming events on the topic
• Tool pages now show which stacks they belong to, so you can go from a product to the job it does
• Each stack carries the date a person last read it against the live products, and states plainly that these are shortlists rather than rankings
• Like the rest of the site, every stack has a Markdown version for AI assistants at the same address with .md on the end

The directory and the job board have both had filters for a while. The events page did not, which made it the odd one out: the only list on the site you had to read top to bottom. It has them now.

What's new?

• /events?cfp=open shows only events whose call for papers is still accepting talks. It is a plain link, so you can bookmark it or send it to someone who is looking for a stage.
• Search across event names, locations, and tags.
• Online or in person, and sorting by soonest event or soonest CFP deadline.
• Every filter lives in the URL. Whatever you are looking at is a link that shows someone else the same thing.
• A CFP countdown on the event rows themselves, so you can see which deadlines are close without switching lists.

Sorting by CFP deadline reorders the list rather than trimming it: events with no deadline move to the bottom instead of disappearing, because a sort control that quietly hides things is worse than one that does not.

The parts of DevRel this site tracks move every week: who is hiring, which tools are worth knowing, which CFPs close before you get round to them. Now you can have that collected and sent to you instead of checking.

What's new?

• A newsletter page that says exactly what an issue contains and how often one goes out
• Signup on the site, with no redirect to a third party and no new tab
• A confirmation step, so nobody can put your address on the list but you
• An unsubscribe page of our own, reachable in one click from the bottom of any issue
• Issues are built from what actually changed here that week, so an empty week does not get an issue

DevRel by the numbers could only ever describe today. Every count on it was recomputed on each build and then discarded, so the one genuinely useful thing about these figures - what they did over a year - had never been recorded. A weekly snapshot now writes them to a permanent record.

What's new?

• An "over time" section on the numbers page, with trend lines for listings, agent-ready listings, open roles, and open CFPs.
• Quarter-over-quarter change, in plain sentences rather than as a dashboard.
• A record that cannot be quietly revised. Rows are added, never edited or removed, and an automated check rejects any change that would rewrite one.

What it deliberately will not do yet

Right now the record holds one row, so most of that section is a promise. It fills in weekly on its own, with no announcement.

Nothing is backfilled. What the directory looked like on a day before the record started cannot be recovered, so it is not going to be estimated.

Comparisons wait for 90 days between two real snapshots, because at this size a week-to-week difference is mostly reshuffling: one company posting four roles moves the job count by a tenth. The salary comparison is stricter again and stays hidden unless both ends of the window have enough disclosing roles to support a median and the move is big enough not to be sample churn.

Each row also records how many employers were being watched that week. Without that, a jump in open roles is ambiguous - the market may have hired, or the board may simply have started watching more companies - and the page says which when the two are tangled.

The worst thing a job board can do is waste your time on a role that closed weeks ago. Every listing here has always been re-read from the company's own board twice a week, and roles that disappear come off. That check was invisible until now.

What's new?

• "Still listed 3 days ago" on job rows and job pages, showing when a run last found the posting on the company's own careers page.
• The stamp travels to AI agents too, through the site's MCP server and the machine-readable listing feed, so a tool answering "is this role still open?" has the same date you do.
• Nominate an employer: if you know a company hiring in DevRel that is not on the board, naming it adds its careers page to the twice-weekly scan. Every DevRel role it posts from then on shows up, not just the one you spotted.

What the stamp does and does not mean

It means the posting was on the company's board that day. That is all it means. Companies leave filled roles up for weeks, so the wording is deliberately "still listed" and never "verified open" - the check cannot tell the difference, so the label does not pretend to.

When a stamp is more than two weeks old it disappears rather than sitting there going stale. A gap usually means a company changed its hiring system or its board stopped responding, and neither is evidence about the role either way. Postings with no stamp at all are simply ones no run has confirmed since this shipped.

Until now the directory could tell you what a tool is, but not how it stacks up against the obvious alternative. That was the gap, and picking a tool is the whole reason people arrive here. So there are now pages that answer the comparison directly.

What's new?

• Head-to-head comparisons. Twenty-four pages comparing pairs of tools we have curated as alternatives to each other, like GitBook vs Mintlify and Discord vs Slack. Both products are measured on identical criteria, so the table actually helps you decide.
• An alternatives page for every listing that has one. Thirty-two of them, such as alternatives to Mintlify. Hand-picked alternatives come first and are labelled as such, with related tools below them.
• DevRel by the numbers. How many tools, communities, and agencies are listed, how many roles are open, how many CFPs are still accepting talks, and exactly how each figure is worked out. Reuse the numbers with attribution.
• Category and tag pages now answer the question. Each one opens with a direct answer, compares the most recently verified entries on the same criteria, and ends with the questions people actually ask.
• Every guide is signed and dated. Documentation pages now show who wrote them and when they were last confirmed accurate, so you can judge how current they are.
• Comparisons are readable by AI tools too. The new pages are listed in llms.txt and each one has a plain Markdown version: add .md to its address.

One thing these pages deliberately do not do is rank a winner or quote prices. We do not test tools hands-on and we do not verify pricing, so every page says what it compares and what it leaves out. Where a product does not publish something, the table says so instead of guessing.

"Is it free?" is the first question anyone asks about a tool, and until now the comparison tables could not answer it. They can now, for 33 of the 47 tools listed.

What's new?

• A pricing row on every comparison, showing whether a tool is free, freemium, paid, or open core.
• Every pricing claim carries the date it was read. Pricing pages change without notice, so a claim with no date is a claim you cannot judge. The tables say "Freemium (checked Jul 26, 2026)" rather than just "Freemium".
• "Open core" means both halves were checked: the product's own source repository carries a recognised open-source licence, and there is a paid tier. A public repository alone does not earn the label.

What we deliberately left blank

Fourteen tools have no pricing shown. For some, the pricing page could not be read; for others, the page genuinely does not say. In both cases the table shows "Not stated" rather than a guess, because a wrong price is worse than no price.

We also do not verify that a published price is what you would actually pay, and we re-check on a schedule rather than continuously. Always confirm on the product's own pricing page before budgeting.

Event pages used to disappear once the event was over. Anyone who had bookmarked a conference, shared it in a group chat, or cited it somewhere was left with a dead link the week after. That is fixed: every event the directory has ever listed keeps its page permanently.

What's new?

• Past events now show a "this edition has ended" page instead of a missing one, with the full record of dates, location, and tags.
• Where the next edition of a conference is already listed, the ended page links straight to it.
• A few July events that had already dropped off the site are back at their original addresses.
• Past events are excluded from search engines, so upcoming events stay the ones people find.

Job pages used to show little more than a title, a location, and a salary range. They now carry the full description from the employer's own posting, so you can tell whether a role is worth your time without opening three tabs. Listings pick their description up the next time their board is checked, so they fill in over the coming days rather than all at once.

Behind the scenes, the board also started keeping records it used to throw away.

What's new?

• Job pages show the full role description, alongside the pay and location already there.
• Roles that state where they hire can now appear in Google's job search results, so more of them get found by the people looking for them.
• Every salary a listing has ever stated is kept permanently, even after the role closes. That history is what a proper DevRel pay reference will be built on.
• The board now records when each role opened and when it closed, so it can eventually show which companies actually hire for DevRel over time.

Checking that a link still loads does not tell you whether anyone is still shipping. Listings that point at a GitHub repository now carry that too, checked weekly rather than taken on trust.

What's new?

• Listing pages show a repository line: star count, the month of the last commit, and the month of the last release where there is one.
• A repository GitHub reports as archived says so on the same line, next to the month the work stopped.
• The figures are coarse on purpose - stars in buckets, dates by month - so the line moves when something real changes, not every week.
• The same facts reach the machine-readable corpus and the get_listing tool on the MCP server, so agents reading the directory see them too.

"How do I get into DevRel" is the most common question in this field and nobody owns a good answer to it. The new Start here page is ours: four steps in order, built from data the site already keeps current rather than advice written once and left to rot.

What's new?

• A reading order through the fundamentals that explains why each guide follows the last, sequenced for someone new rather than mirroring the documentation menu.
• The communities, newsletters, and podcasts that an automated check confirms are still reachable, with the date each was verified. Dead invite links are the norm in this space.
• The calls for papers closing soonest, with a calendar you can subscribe to.
• An honest read of the job market, including how many of the open roles are actually junior. Right now that number is zero, and the page says so.

Listings now carry structured facts instead of prose alone. Where a tool publishes an llms.txt file or an MCP endpoint, the directory records it - verified by the same weekly health check that keeps listings alive, not self-reported.

What's new?

• A new agent-ready view lists every tool an AI agent can consume directly.
• Listing pages show a spec block: repository, documentation, and machine-access links where they exist.
• The same facts flow into the machine-readable corpus, so agents reading the directory learn them too.

Tags across the directory are now navigation, not just labels. Click a tag on any listing to see everything else that shares it.

What's new?

• Tags on directory rows and listing pages now link to their tag pages.
• Searching the directory updates the address bar as you type, so a filtered view can be shared or bookmarked.
• Tag links pick up the same keyboard focus styling as the rest of the site.

The directory has been strongest on tools. The agency, newsletter, and podcast categories existed, but each held a single listing. That changes today.

What's new?

• Agencies: technical content and developer marketing shops, from solo consultancies to teams serving large platforms, for when you need outside help reaching developers
• Newsletters: regular reads covering new devtools and how engineers build and grow products
• Podcasts: interview shows with devtool founders, open-source company builders, and the people who make developer tools

Every link was fetched and checked before listing, newsletters and podcasts included only if they are actively publishing. As always, each description says what the thing is in plain terms, not the vendor pitch. Know an agency, newsletter, or podcast that belongs here? Submit it from the directory page.

You can now follow the whole site without visiting it. Three new RSS feeds cover the jobs board, the events calendar, and the directory, alongside the existing blog feed.

What's new?

• Jobs feed: every open role as it lands, newest first, with the salary in the title whenever the posting discloses one.
• Events feed: newly added events with open CFPs flagged in the title, so speaking deadlines show up in your reader on time.
• Directory feed: the latest tools, communities, newsletters, podcasts, and agencies added to the directory.
• New this week: one page with everything added in the last 7 days across listings, jobs, events, posts, and site updates.
• Your feed reader can discover all of these on its own: the relevant pages now advertise their feeds.

AI agents and developers now have one canonical page describing how to read this site programmatically.

What's new?

• For agents and developers: the Markdown mirrors, the MCP server, the feeds, and the fair-use rules, in one place.
• robots.txt now welcomes the major AI crawlers by name, with the payment endpoints excluded for every crawler.

Hiring for Developer Relations now has its own front door. The new hire page explains how to post a role: community listings are free by email, and a featured post is pinned above the list in a highlighted card for 30 days.

What's new?

• One page with everything an employer needs: what a featured post includes, live board numbers, and how to get in touch.
• The board's hygiene rules, spelled out: listings expire automatically, live roles are re-checked twice a week, and salaries are shown exactly as posted.
• After checkout you now land back on the page you started from.

Trust in a directory comes from knowing how it is maintained. Those rules are now a page you can read instead of something you have to take on faith.

What's new?

• How we curate states the whole process: what qualifies for a listing, how every entry keeps being re-checked after it goes live, when a listing is removed, and why paid placement can never touch editorial judgment.
• The Verified badge on a listing page now links to that page, so "verified" has a definition one click away.
• llms.txt and llms-full.txt now open with a "Last updated" date, and the MCP server returns the same date with listing details and directory stats, so AI tools can tell how current the data is.

Until now, jobs on the board linked straight out to the posting, so there was no way to share or bookmark one. Every listing now has its own page: salary exactly as the posting states it, location and working arrangement, tags, and the apply link front and center.

What's new?

• Job rows grew a details link; clicking the row still opens the posting directly, like before.
• Shared job links show a proper preview card, and site search now lands on the job's page.
• A listing that closes keeps its page as a clearly marked record that points to the company's remaining open roles and back to the board.
• For AI agents: job results from the MCP server now cite these pages instead of a board anchor, so citations survive the listing.

The directory used to be a single filtered page. Now every slice of it has its own address you can bookmark, share, and find through search engines.

What's new?

• Category pages like tools and communities for each of the five listing categories.
• A new tool sub-taxonomy with six browsable types, from docs & dev portals to analytics & listening.
• Tag pages for every tag with enough listings behind it, linked from the tag chips on each listing's page.
• The category pills on the directory now lead to these pages, and search keeps working as before.

You can now subscribe to the events page from your own calendar. Two feeds are available: an events calendar with every upcoming conference as an all-day entry, and a CFP deadline calendar with one entry per open call for papers.

What's new?

• Subscribe controls on the events page, with one-click subscription links and plain .ics URLs for any calendar app.
• The CFP calendar carries a reminder one week before each deadline. Apple Calendar, Outlook, and Thunderbird fire it; Google Calendar shows the entries but not the reminders.
• Past events and closed CFPs drop out of the feeds automatically.

Until now, events on the events page linked straight out to the event site, so there was no way to share or bookmark one. Every upcoming event now has its own page with the dates, location, tags, CFP status and deadline, and a link to the organizer's directory listing when we have one.

What's new?

• Event rows now open the event's page; the external site stays one click away.
• Shared event links show a proper preview card, and search results land on the event page directly.
• For AI agents: event results from the MCP server and site search now point at these pages instead of the events list. Pages exist while an event is upcoming and lapse afterwards.

Site search already covered the docs, blog, directory, and jobs. Now it also finds events: search for a conference and it shows up in the results, linking straight to that event on the calendar.

What's new?

• Search for a conference name and it appears in the results, deep-linking to the event
• Events are matched by name and location, so city searches work too
• Conferences whose call for papers is still open are marked "(CFP open)" so you can spot speaking opportunities at a glance

The jobs page now opens with a snapshot of the DevRel job market, computed live from the board itself. AI agents get the same numbers in one call through a new MCP tool.

What's new?

• A market band on /jobs: open roles, the share that are remote or remote-friendly, how many companies are hiring, and the workplace mix.
• A median salary headline joins the band once enough postings state USD pay. Small samples make misleading headlines, so we hide the figure rather than caveat it.
• A sixth MCP server tool, directory_stats, returns aggregate counts for listings, jobs (including the remote/hybrid/onsite mix), and events in one call. Salary figures always carry the sample size, and currencies are never blended.

Listing pages used to show little more than the same one-liner you already saw in the directory overview. Now every entry has a proper write-up.

What's new?

• Every directory listing carries a short researched summary: what the product actually does, who it is for, and which part of DevRel work it serves.
• The write-ups reflect the current state of each product; we verified every claim against the vendors and noted recent pivots, rebrands, and acquisitions where they matter.
• AI agents get the same depth: fetching a listing through the MCP server now returns the full write-up alongside the basics.

Directory listings are no longer a flat list. We hand-curated relationships between tools, so a listing now shows genuine alternatives and the tools it pairs well with, alongside the existing tag-based similar listings.

What's new?

• "Alternatives" and "Works well with" sections on listing pages, starting with the docs, community, analytics, and events clusters.
• Relationships are curated by hand, not computed - an entry only appears when we would actually recommend considering it.
• For AI agents: the MCP server's get_listing tool now returns these relationships too.

The site's machine-readable surface now covers the data, not just the docs. Point an AI tool at devrel.directory and it can read the whole directory without clicking through pages.

What's new?

• llms-full.txt now includes every directory listing, every active job, and every upcoming event and CFP alongside the docs, and it always reflects the current data.
• Every listing has a Markdown twin at /directory/<slug>.md - the same content as the listing page, in plain Markdown for agents and scripts.
• Listing pages grew the same "Copy Markdown" and "Open in ChatGPT / Claude / Cursor" controls the docs pages already have.
• llms.txt points to all of it.

AI agents connected to the MCP server can now filter open roles by seniority level and minimum salary, on top of the existing query, remote, and employment-type filters. Agents and the jobs page now share the exact same filter logic, so an agent's answer always matches what you see on the site.

What's new?

• The list_jobs tool accepts a seniority level (junior to leadership) and a salary floor.
• Salary filtering works the way the site's filter does: a role qualifies on the highest pay its posting states, and roles without stated pay never qualify.

The MCP server is now listed in the official MCP registry as directory.devrel/devrel-directory, with domain ownership verified.

Registry-backed clients and catalogs can discover the server by name and connect to https://devrel.directory/mcp directly. Nothing changes for existing connections.

Two additions this week: every directory listing now has a proper home on the site, and AI agents can plug into devrel.directory directly.

What's new?

• Every directory listing now has its own page with its description, tags, and a set of similar tools from the same category, so you can link to and share individual listings.
• devrel.directory now runs a read-only MCP server, so AI agents like Claude can search the directory, jobs, events, and docs and cite the site as the source. Connecting takes one line and no signup.

The tools category of the directory grew from 9 listings to 49. Every addition was researched across community catalogs, curated lists, and the current landscape, then held to one test: would a DevRel practitioner actually reach for this in their day job?

What's new?

• Docs tooling from hosting to style linting: GitBook, Read the Docs, Material for MkDocs, Starlight, Vale, and Algolia DocSearch
• The AI-era stack: assistants that answer developer questions from your docs, and tools that track how your product shows up in AI answers
• API and SDK tooling: the generators and reference platforms behind some of the best-known developer APIs
• Ways to measure DevRel: open-source adoption analytics, GitHub trend benchmarking, and privacy-friendly web analytics
• Community, events, and speaking: chapter programs, CFP management, conference discovery, and live audience Q&A
• Content production: streaming, terminal recording, developer slide decks, newsletters, and technical blogging

Every description says what the tool does for a DevRel team, not the vendor pitch, and every link was checked before it went in. Know a tool that belongs here? Submit it from the directory page.

Getting something listed no longer requires a GitHub account. The submit buttons on the directory, jobs, and events pages now open a pre-filled email with exactly the fields we need to review your submission.

What's new?

• Submit buttons across the site open a ready-to-send email - no account needed
• The contributing guide is rewritten around the new flow
• The Contribute link in the navigation now leads straight to that guide
• Blog comments have been retired

The jobs board grew from a simple list into a real tool for finding your next DevRel role.

What's new?

• Search across titles, companies, locations, and tags.
• Filters for remote-only roles, employment type, seniority, and minimum salary - and filtered views are shareable via the URL.
• Salary ranges shown exactly as the postings state them, never estimated.
• Posting age at a glance, plus a "New" marker on roles from the last week.
• Featured roles now get a proper spotlight: pinned cards above the list with everything visible at a glance.

Featured posts on the jobs board can now be purchased directly through Stripe Checkout. Pick a brand-new listing or an upgrade of an existing one, pay, and the rest is automated: the paid listing is prepared for publication and goes live after a quick editorial review.

What's new?

• Self-serve checkout for featured job posts, handled by Stripe as the merchant of record
• Works for new listings and for upgrading a listing that is already on the board
• A featured slot pins the listing to the top of the jobs board for a month
• Every purchase still passes a human review before it appears, so the board stays curated
• The email path remains available if you prefer to arrange it directly

DevRel events now have a home here: the events calendar lists upcoming conferences and community events alongside CFPs that are still open.

What's new?

• Browse upcoming DevRel-relevant events, each linking straight to the official site
• See which CFPs are still open, sorted by deadline
• Know an event we're missing? Submit it through the new event submission form

The four core fundamentals pages were rewritten from generic overviews into reference material.

What's new?

• Definition of Developer Relations: a working definition plus clear positions on what DevRel is not and where it should report
• The importance of DevRel: the business case, with the network-effects mechanism and a worked example from a public filing
• Roles and responsibilities: modern role list, who to hire first by company stage, and reporting-line tradeoffs
• Understanding developer personas: a persona template, a fully worked example, and research methods with honest effort guidance

The homepage now shows what is new instead of staying the same between visits.

What's new?

• A "Latest insights" section on the homepage with the three newest posts from the blog
• The announcement banner at the top of every page now links to the most recent article, so it stays current as new posts are published
• Dismissing the banner hides it until the next new article comes out

The homepage has a new look. It leans into a terminal aesthetic: a console-style panel shows live counts for every part of the site, and each section is now one click away from an indexed list.

What's new?

• A console-style overview panel with real counts for docs, blog, directory, events, jobs, and changelog
• A full site index so every section is reachable straight from the homepage
• A short section on why this site exists
• Latest blog posts in a dated list with one-line summaries
• A one-click button to copy the whole knowledge base as llms.txt for your AI assistant
• A fresh lime accent color throughout the site

DevRel roles now have a home here: the jobs board lists open positions in developer advocacy, community, docs, and DevRel leadership.

What's new?

• Browse open DevRel roles, each linking straight to the official posting
• Listings stay live for 60 days, newest first
• Hiring? Community listings are free, and featured posts get pinned on top

You can now follow the site from anywhere, and search engines can find every page.

What's new?

• An RSS feed with every blog post, so new articles land straight in your feed reader
• A new RSS link in the footer, next to GitHub
• A sitemap covering the docs, blog, changelog, directory, and jobs pages, so search engines index everything

Links you share from devrel.directory now come with a proper preview image.

What's new?

• Every page generates its own preview card, showing the page title on the site's dark theme
• Cards appear wherever link previews are supported: social platforms, chat apps, and messengers
• Docs, blog posts, the directory, and the jobs board are all covered

The DevRel strategy page is now a quick-reference framework instead of a second long read.

What's new?

• The framework in four parts: goals, audience, activities, and metrics, each with concrete examples
• The three failure modes that sink most DevRel programs, each with a self-test
• A pointer to the comprehensive blog guide for the full walkthrough

You can now suggest additions to the site through structured submission forms. The Submit buttons on the directory lead straight to them.

What's new?

• Suggest a tool, community, newsletter, podcast, or agency for the directory
• Submit an open DevRel role for a free listing
• Every submission is reviewed before it goes live

Site search used to look through the docs only. Now it covers everything: blog posts, directory listings, and open jobs show up in the results alongside the docs, each labeled by section.

What's new?

• Search for a topic and matching blog posts appear next to the docs
• Tools, communities, newsletters, and podcasts from the directory are searchable by name, description, and tags
• Open DevRel roles turn up when you search for a job title or company

Developers increasingly arrive through AI assistants and coding agents, not browsers. The new AI-era DevRel section is a practical, first-principles guide to that shift.

What's new?

• Agents Are Your New Developers - what changes when first contact with your product is a machine, and how to audit your quickstart with an agent
• Writing Docs LLMs Can Use - llms.txt, Markdown mirrors, and the structure that makes docs retrievable, with this site as the live example
• MCP Servers Are the New SDK - treating an MCP server as a developer relations surface
• Measuring Adoption You Cannot See - honest instruments for AI-mediated adoption

The section sits right after Fundamentals in the docs sidebar.

devrel.directory now practices the AI-era DevRel it writes about. AI assistants and crawlers can read the whole knowledge base in plain Markdown, and every docs page has quick actions for working with it in your AI tool of choice.

What's new?

• /llms.txt - a Markdown index of the docs and blog for AI agents, following the llms.txt convention
• /llms-full.txt - the entire documentation as one Markdown file
• Every docs page is also available as raw Markdown by appending .md to its URL
• New actions on each docs page: copy the page as Markdown, view it raw, or open it directly in ChatGPT, Claude, or Cursor

Two highlighted callouts in our DevRel strategy guide still used bright blue boxes from before the site's design refresh. They now use the same card style as the rest of the site.

What's new?

• Highlight boxes in the DevRel strategy guide now match the site-wide design
• They read comfortably in both light and dark mode

The icons in our blog posts still carried their old mixed colors from before the site's design refresh. They now use the same calm, unified style as the rest of the site.

What's new?

• Icons across all blog posts now match the site-wide design
• Less visual noise, so the content itself stands out more

The changelog you are reading right now got a layout upgrade. Updates now sit on a timeline, so it is easier to scan what changed and when.

What's new?

• Each update shows its date and tags in a column next to the entry
• Entry titles and text follow one consistent style
• Updates always appear newest first

The site now does what its name promises: there is an actual directory at /directory. It is a curated map of the DevRel ecosystem, seeded with the resources practitioners actually use.

What's new?

• Browse listings across five categories: tools, communities, newsletters, podcasts, and agencies
• Filter by category or search across names, descriptions, and tags
• Every listing links straight to the resource itself
• Know something that belongs here? Submit a listing right from the page

We removed the placeholder pages that were sitting in the docs navigation without real content. Everything you see in the sidebar now is an actual article, and the start page reflects what the knowledge base currently covers.

What changed?

• Placeholder pages were removed from the docs and the navigation
• The start page was rewritten to point at the content that exists
• Small fixes across the fundamentals pages (typos and formatting)

More sections will come back as they are written properly - follow the changelog or the blog for updates.

When you shared a devrel.directory link in chat apps or on social platforms, the preview card could show the wrong site name. Link previews now show DevRel Directory with the right page title and description.

We fixed an issue where documentation and blog pages could display without their intended text styling. Headings, lists, and article layouts now render the way they were designed to.

What changed?

• Documentation pages show proper headings, spacing, and typography again
• Blog posts got their formatting back, including the table of contents and article layout

devrel.directory has a new visual identity. The site is now dark by default, with a single blue accent and a calmer, more focused layout across every page.

What's new?

• A redesigned start page that shows what lives here: the docs, the blog, and the changelog
• Dark mode is the default look, and the light theme is one click away
• The blog index is easier to scan, and posts got a cleaner header
• The changelog you are reading right now was restyled to match

The knowledge base now has a Practices section with three in-depth guides.

What's new?

• DevRel Activities - the core activity clusters, when each one earns its place, and how to combine a few into a coherent program
• Community Building - starting, growing, and measuring a developer community, plus how to show up well in communities you do not own
• Content Creation - content types and their jobs, writing craft for developers, distribution, and a sustainable pipeline

These replace pages that previously sat in the navigation as placeholders. More sections are being expanded the same way - follow the changelog for updates.

We've added a changelog to the site. This is the first entry.

What's new?

We're excited to introduce our new changelog feature! This addition will help keep you informed about the latest updates, improvements, and features added to the devrel.directory.

Keep an eye on this space for future updates!

Future iterations will also include credits to the contributors who made the changes. 🤗