Fractional DevRel proposal · v1 · May 27, 2026

A 12-month DevRel program engineered to make Coral the category-defining ADAP.

An outside-in plan for Coral — the open-source, local-first SQL runtime for AI agents — built on the same operating system we've run for Memori, Glean, CrewAI, and Perplexity. Tuned to Coral's wedge: token efficiency, MCP-native architecture, Apache 2.0, and the upcoming Coral Cloud launch.

Engagement: Fractional, 90-day pilot → 12-month North star: Weekly active devs running ≥3 sources Cloud launch: Q3 2026
The setup

Coral has the strongest architectural thesis in the ADAP market. Architectural lead is perishable.

A 42-vendor SWOT (Steampipe, MindsDB, CData, Composio, PromptQL, Zapier MCP, vendor MCPs from Datadog/Honeycomb/Grafana, AI SRE agents, IDPs, federation engines) confirms the wedge — and the timer.

4.9k
GitHub stars
5 weeks post-launch · top 1% velocity
102
Sources today
27 bundled + 75+ community
3.4×
Cost-efficient vs MCPs
Claude Opus 4.6, n=82 tasks
0
Funding announced
Crunchbase, May 2026

Why DevRel is the lever. Coral does not need more engineering hours — it needs community-led plugin velocity, public reference customers, and a defining content narrative before Steampipe relicenses, MindsDB pivots agent-native, or Anthropic ships a first-party MCP aggregator. We attack those three gaps directly.

The 7 initiatives

Seven coordinated programs, each attached to a SWOT lever.

Same 7-initiative framework we ran for Memori and CrewAI, tuned to Coral's wedge. Every program ties to a specific Strength to amplify, Weakness to close, Opportunity to capture, or Threat to neutralize.

1

Source Spec Marketplace

Attacks W1 (integration breadth), O4 (no dominant OSS leader)

Public registry at withcoral.com/sources, tiered badges (Official / Verified / Community), Source Spec Cookbook, automated linting, weekly author office hours, featured-source rotation.

2

BYO Agent Reference Architectures

Attacks S8 (MCP-native), O5 (composability with code-context tools)

Official reference repos + 5-min videos + token-savings calculator for Claude Code, Cursor, Continue.dev, Sourcegraph Cody, Goose, Aider, Devin, Cline. Co-marketed with Anthropic for the Claude Code reference.

3

Token Efficiency Benchmark Series

Attacks S2 (defensible performance claim), T3 + T8 (narrative risk)

Quarterly published benchmarks across new model releases (Opus 4.7, GPT-5, Gemini 3) on identical task sets. Open-sourced harness. Side-by-side vs Composio, Anthropic reference MCPs, Steampipe, vendor MCPs, MindsDB. We define the metric before competitors do.

4

Coral Co-Pilot Cohort (Design Partners)

Attacks W3 (no customers), W2 (no commercial model — pre-sells Cloud)

12-week cohort, 8–10 companies per batch, 4 batches/year. White-glove from the founders. Free Cloud access for life at 50% off. In exchange: case study, quote, co-presented talk. Public Demo Day livestreams each quarter.

5

Source Spec Bounty Program

Attacks W1 (integration breadth)

$200–$2,000 per merged spec depending on complexity tier. $5K top quarterly contributor, $10K annual grand prize. Public leaderboard. Quarterly themed Hack Weeks (Observability, Customer-Data, Security-Tooling, Finance). $150K/yr budget targets ~100 new specs/year.

6

Community + Content Engine

Attacks W3, W8 (brand confusion)

Weekly office hours (Tuesdays 11am PT). Discord investment (7 dedicated channels). 1 team post + 1 community post per week. Biweekly newsletter The Coral Reef. Weekly YouTube. Podcast guesting (Latent Space, AI Engineer, Practical AI, ChangeLog).

7

Strategic Channel Integrations

Attacks T6 (IDP displacement), T5 (Composio commoditization)

Backstage plugin (Apache 2.0), Cursor Marketplace featured entry, VS Code MCP gallery, Smithery/mcp.run/Klavis cross-listing, AWS + Azure Marketplace for Cloud, Anthropic co-marketed reference architecture.

12-month phasing

Three phases, three marketing moments. Cloud launches in Phase 2.

PHASE 1 · MONTHS 1–3 · JUN–AUG 2026

Foundations

Source Spec Cookbook v1 · Reference architectures for Claude Code + Cursor · Benchmark drop on Opus 4.7 · Discord investment, weekly office hours · Cohort #1 applications open · Bounty program launches with $50K initial pool.

Milestone: 8k stars, 130 sources, 1,500 weekly active devs.

PHASE 2 · MONTHS 4–6 · SEP–NOV 2026

Coral Cloud / Teams GA

Cloud launches with multi-tenant SaaS, shared credential vault, audit logs, RBAC. AWS Marketplace listing. Backstage plugin GA. Cohort #1 Demo Day livestream. Cohort #2 opens. Second benchmark drop. Anthropic co-marketed launch.

Milestone: 12k stars, 180 sources, 3,000 WAD, 25 Cloud teams paying.

PHASE 3 · MONTHS 7–12 · DEC 2026 – MAY 2027

Scale

AI Engineer World's Fair + KubeCon presence. Plugin SDK v2. Cohorts #3 + #4. Third + fourth benchmark drops. First analyst recognition targets (Gartner Cool Vendor, Forrester Wave inclusion).

Milestone: 20k stars, 250 sources, 5,000 WAD, 50 Cloud teams, $500K ARR.

What we deliver

Fractional engagement. The operating system, not a hire.

Coral keeps engineering focus on the runtime. We bring the playbook, the tools, the talent network, and the cadence.

Always-on operating system

  • DevRel strategy ownership + quarterly OKRs
  • Content engine: editing, writing, video production
  • Community ops: Discord, office hours, newsletter
  • Benchmark publishing cadence + reproducibility infra
  • Cohort program management end-to-end
  • Bounty triage + review queue management
  • Conference talk authoring + submission machinery
  • Analyst + press outreach
  • Brand / SEO operations
  • Monthly metrics dashboard + executive readout

Project-based deliverables

  • 5–7 BYO Agent reference architectures
  • Source Spec Cookbook + Plugin SDK docs
  • Coral Cloud / Teams GTM (Phase 2)
  • Backstage plugin + other channel integrations
  • 8–10 cohort case studies per cohort
  • Quarterly benchmark publications

Tooling we deploy

  • OpenClaw — internal AI DevRel agent that triages community questions, drafts outreach, routes leads
  • MetricsBoard — public metrics dashboard
  • Cohort Portal — application + onboarding flow
  • Bounty Tracker — public leaderboard + payout pipeline

Engagement model

  • 90-day pilot: $35K/mo retainer + $50K program seed = $155K total
  • Year 1 full: $420K retainer + $460K programs + $40K tools = $920K
  • Quarterly OKR review gate — scope expands or contracts based on outcomes
  • Exit ramp: at month 12 we hand off the operating system to a hired Head of DevRel
Why opencolin

Four prior programs, one operating system.

We've run this same 7-initiative framework end-to-end at four other AI/ML companies. Coral inherits the playbook on day one.

AI workflow

CrewAI

Interactive strategy site with ecosystem, community, marketplace, and partner programs. Same 7-initiative framework adapted to the multi-agent orchestration market.

opencolin/crewai-devrel →

Agent memory

Memori

SF startup residency, weekly workshops, monthly hackathons, bounty system, grants, MCP marketplace. The pattern we're proposing for Coral.

opencolin/memori-devrel →

Live web search

Perplexity Sonar

90-day growth plan and ecosystem vision for the Sonar API. Strategic presentation focused on developer acquisition and use-case expansion.

opencolin/perplexity-devrel →

DevRel infra

AI DevRel OS

Our internal operating system for running fractional DevRel — Next.js app with OpenClaw agent, PostHog tracking, CRM integration. Powers all four programs.

opencolin/ai-devrel →

Next steps

Let's start with a 90-day pilot.

$155K covers Phase 1 Foundations — Source Spec Cookbook, weekly office hours, Cohort #1 prep, first benchmark drop, two reference architectures, Discord investment. Quarterly review gate to expand into the full Year 1 engagement.

Email collin@dabl.club Walk through the deck