Content

Own the "agent era runs on compute you can trust" narrative before incumbents define it.

Tenki's blog is empty. That's the opportunity: a publishing flywheel where benchmarks, reference architectures, migration teardowns, and build-in-public all reference each other. One team deep post every two weeks, one community post weekly, a biweekly newsletter, and weekly YouTube — all anchored to the trust-and-honesty wedge.

First 90 days

12-week content calendar.

Each week ships one major artifact. Sequencing ensures each launch rides the air cover of the prior week's audience — opening with a migration win and the Benchmark Drop #1 teardown.

WkDropChannelPersona
1"I migrated a real monorepo to Tenki in 2 minutes — here's the diff and the bill"Blog + XCI owners
2Benchmark Drop #1: AI reviewers on real OSS bugs (the teardown)Blog + HN + XEng leads
3"Give your agent root without giving it yours" — Sandbox + Claude CodeYouTube + BlogAgent builders
4Tenki vs GitHub Actions vs Depot vs Blacksmith (honest comparison + calculator)Blog (SEO)Evaluators
5Cohort #1 applications openLanding + email (dabl.club)AI-native teams
6"The precision problem in AI code review" (POV essay)Blog + XEng leads
7Codex + Tenki Sandbox reference architectureGitHub + BlogAgent builders
8Community story #1 ("How I ship with Tenki")Blog + LinkedInAll
9Weekly stream milestone: live-fix an OSS bug w/ TenkiYouTube/X liveAll
10Customer story #1 (cohort reveal)Blog + LinkedInEnterprise
11"Runners, review, sandbox: why they're one product" (the convergence thesis)Blog + HNEng leads
12Agent Camp × Tenki hackathon announceLanding + email + XAgent builders
Content pillars

Four pillars, one narrative.

Every artifact serves one of four pillars — and all four ladder up to a single through-line: the agent era runs on compute you can trust, and Tenki is where it runs.

Pillar 1

Honest benchmarks & trust

Reproducible, open-harness benchmarks across AI review, runners, and sandbox — published even where Tenki loses (precision today, the long-running-agent frontier next). The honesty is the credibility, and it's how you win developers: be useful, build in public, let the receipts do the selling. Tenki defines the metric.

Cadence: quarterly + per-model-release. See Benchmarks.

Pillar 2

Long-running agents & sandbox-first

Own the category no one's claiming: the sandbox for long-running, self-improving agents (incumbents optimize short-lived sandboxes). Canonical reference architectures — Claude Code, Codex, Cursor, Devin, Cline on Tenki Sandbox — plus "give your agent root without giving it yours." Repo + 5-min video + guide + cost calculator each.

Cadence: ~1/month in year 1.

Pillar 3

Migration & cost

Migration Wizard storytelling, honest comparison guides vs GitHub Actions / Depot / Blacksmith / Namespace, and the public cost calculator. Lean into the GitHub-pricing moment.

Cadence: evergreen SEO + reactive to pricing news.

Pillar 4

Build-in-public

The weekly stream running Tenki against a live OSS bug, plus community "How I ship with Tenki" posts and OSS PRs with diff + prompt history attached — the exact motion the JD describes.

Cadence: weekly stream + 1 community post/week.

Cadence

The publishing drumbeat.

A sustainable, predictable rhythm — heavy enough to build a brand from zero, light enough for a lean DevRel team to actually ship every week.

StreamCadenceOwner
Tenki-team deep post1 every 2 weeksDevRel + eng
"How I ship with Tenki" community post1 per weekContributor-led, lightly edited
Newsletter (biweekly)Every other TuesdayDevRel
YouTube (migration + Sandbox walkthroughs)Weekly · ThursdaysDevRel
Build-in-public stream (live OSS bug)Weekly · FridaysDevRel
Coordinated HN / Reddit / X launchesPer major dropDevRel

Seeded distribution. Every drop gets cross-pollinated to dabl.club's 86k developers (attributed via UTMs) — the instant top-of-funnel most startups spend a year building. Newsletter target: 8,000 subscribers by month 12.

YouTube + podcasts

Show, don't tell.

Video is where migrations and agent setups become believable. Podcast guesting is how a founder-DevRel borrows existing audiences while Tenki's own brand compounds.

YouTube cadence

Podcast guesting targets (year 1)

Topic doors we offer hosts

  1. "The precision problem in AI code review — and why we publish where we lose"
  2. "Runners, review, and sandbox are secretly one product"
  3. "Compute is the new digital oil: building dev tools on owned metal"
  4. "Give your agent root without giving it yours: microVMs vs containers"
  5. "What the GitHub Actions pricing shift means for CI economics"
SEO + brand

Own the evaluator's search.

When a team is deciding whether to switch CI or adopt an agent sandbox, they Google a comparison. We make sure Tenki's honest comparison is the result they find — and we own the agent-era category terms before the pure-plays do.

Target queries (rank #1 by month 12)

Brand hygiene checklist

How this rolls up. The content engine attacks W3 (no content engine / empty blog) and feeds every other initiative: benchmarks build the trust that converts trials, reference architectures are the agent-era portfolio, and comparison SEO captures O1 (the GitHub-pricing backlash). Tracked weekly on the metrics dashboard.