Bezel Team strategy-synthesis strategy-001
strategy-synthesis artifact · for Bezel Team · phase strategy-001 · status draft
strategy-synthesis artifact · for Bezel Team · phase strategy-001 · status draft
No explicit evidence field yet. Require tests, screenshots, linked PRs, or reviewed outputs before marking complete.
- kind: strategy-synthesis
- version: 1
- created at: 2026-05-04T12:02:39.018Z
- project id: bezel-team
- first 30 days: 4 items
- pricing stance: why_not_seats: string, recommendation: string, why_not_tokens: string, initial_test_prices: object, why_not_pure_per_run: string
- model comparison: claude: object, gemini: object, atlas_codex: object
- source artifact id: e89f24e2-a12d-4b3f-b96e-aee2c8e6e83e
- company product rule: company: string, product: string, naming_rule: string, paid_product: string, public_wedge_later: string
- recommended direction: v1_loop: string, category: string, one_liner: string, first_buyer: string, first_wedge: string, v1_do_not_build: object, ai_first_reframe: string
- next artifact recommendation: Create INTENT.json next, explicitly locking category, buyer, wedge, V1 loop, pricing hypothesis, and non-goals before architecture.
Machine-readable source fields
strategy-synthesis
1
2026-05-04T12:02:39.018Z
bezel-team
- Interview 10 founder-led or 5-15 person engineering teams; inspect their actual backlog and identify chore/debt issues they would delegate.
- Build the thinnest hosted Team loop against GitHub only: label issue, run worker, open PR, show status/proof.
- Run Bezel on Dreamborn/Bezel internal repos first, then one design partner repo with supervised setup.
- Success gate: one real external customer merges at least five Bezel-generated PRs and says they would pay for more capacity.
AI-first value is autonomous throughput, not number of human supervisors.
Use capacity packages based on completed/review-ready task units for the buyer-facing model; meter agent-hours internally for margin and abuse control.
Tokens expose implementation detail and make Dreamborn look like a model reseller.
| price | package | purpose | included |
|---|---|---|---|
| $499/mo | Design Partner | validate willingness to pay and usage shape | up to 100 review-ready task units or agreed fair-use cap |
| $1,499-$1,999/mo | Team | first real commercial package | larger task capacity, hosted dashboard, managed workers, persistent history |
Punishes retries and messy real engineering work.
Backlog debt relief for 5-15 person engineering teams.
Bezel Core, ledger, multi-provider support, and enterprise features are distractions before paid Team validation.
Bezel should be work management for AI coding agents, not an infrastructure control plane.
Hybrid active-repo base plus included agent-hours and overage.
Chore backlog for founder-led/small engineering teams.
Do not build Core or custom ledger before one real team merges one Bezel-generated PR.
AI-first framing should be autonomous capacity/team member, not DevOps tooling.
Capacity subscription around Task Units from issue intake to review-ready or merged PR.
A small team connects GitHub, labels chore/debt issues, Bezel returns PRs with proof bundles for human review.
If the first screen feels like logs, pipelines, ledger, or agent configuration, the product has missed the AI-first shift.
The category is managed autonomous engineering work. The first product should sell delegated throughput with reviewable accountability.
Start with simple task-capacity packages for buyer clarity, track internal cost by agent-hour, and reserve repo/seat concepts for limits not value.
e89f24e2-a12d-4b3f-b96e-aee2c8e6e83e
Dreamborn
Bezel
Dreamborn is the company. Bezel is the product. RedKey is legacy/internal only.
Bezel Team
Bezel Core
GitHub issue label -> Bezel queue -> isolated worktree -> Codex run -> proof bundle -> PR -> human review -> merge/decline -> run history.
Work management for autonomous coding agents
Assign your backlog to Bezel. Review the PRs. Merge what is good.
Founder-led or 5-15 person engineering teams with real backlog debt and velocity pressure.
Chore/backlog debt relief: dependency updates, tests, docs, small migrations, low-risk bug fixes, hygiene tasks.
- Bezel Core CLI
- Linear integration
- multi-agent-provider matrix
- enterprise SSO/RBAC
- custom public ledger story
- marketplace
- autonomous merge without explicit policy
Do not price or design around humans using software. Price and design around autonomous work capacity that humans supervise.
Create INTENT.json next, explicitly locking category, buyer, wedge, V1 loop, pricing hypothesis, and non-goals before architecture.