Free runs on your machine
init, sync, doctor, and core MCP tools stay local. No card required to try a real repo.
// governance + audit layer for AI coding agents
[ repo-bound memory ] · build v4.5.3
one rulebook across every CLI · every run leaves a receipt · your source stays local
your1/7agents
stopre-explainingstarting
from0.
0dai keeps the roadmap, rules, and review receipts in your repo's ai/directory — so every CLI works under the same rules and every run leaves a receipt you can replay.6 terminal CLIs + Cursor MCP bridge
projects initialized
terminal CLIs
registered MCP tools
// one source · terminal CLIs + editor bridge
Claude CodeCodexGeminiAiderOpenCodeQoderCursor MCP bridge
Works with your favorite AI coding agents
Claude Code
Codex
Gemini
Aider
OpenCode
Qoder
// note from the builder
I built 0dai because I was tired of explaining my repo for the 47th time. Every new chat, every new agent, the same setup tax. The agents were not dumb — they were amnesiac. So I made the memory layer they should have had.
shipping log6 entries · 8 months
- sprint 1
47th time explaining the repo. Ragequit. Opened editor.
- sprint 2-3
Safety rails + ai/ scaffold. First dogfood task shipped.
- sprint 4-5
Working group v2.2 · 6 deliberation profiles.
- sprint 6
Ghost peer-critic · 2-of-3 consensus. Race caught before merge.
- sprint 7
700+ tests. v4 CLI. Friday digest cadence running.
- shipping in public
Dogfooded on its own repo, every change with receipts.
for solo founders
Know what you pay. Keep what you build.
You ship alone. You should see the bill before checkout and own the repo if 0dai stops tomorrow. No black-box memory, no hostage data.
Pro is $15/mo — listed before you click
Seat price and token estimates live on the pricing page with a calculator. No surprise invoice after a trial.
Export is a command, not a ticket
0dai export --all writes a tarball you can inspect. Today it includes personas, path-protect policy, and usage ledger when present; remaining surfaces are labeled in progress.
Your repo outlives the vendor
- Memory lives in ai/ as YAML and Markdown — grep works, git history applies.
- Switch agent CLIs without rewriting project context; the same files drive every one.
- Self-host the open core or walk away with files any agent can read tomorrow.
two ways to run agents.
only one of them compounds.
today
- half your AI bill goes to re-explaining the repo
- every chat starts from zero context
- the refactor that failed last week gets tried again
- no idea which model is good at which task type
with 0dai
- context delivered, not re-paid for
- projects remember decisions, outcomes, patterns
- failed paths flagged before agents repeat them
- model routing based on what actually shipped
Not a fit for every team. See where Claude Code subagents beat 0dai — the honest feature matrix.
the moat is curation, not access.
raw graph access would commoditize the product. subscribers get curated bulletins and routing data instead.
A · repo-resident
free. ships in every project.the scaffold
safety rules, governance, agent-config templates, master-plan scaffold, onboarding ritual. Installs into your repo and stays.
- works offline
- lives in ai/ next to your code
- yours to keep, no cloud required
B · cloud-only
the network. subscribers only.the curation
cross-project pattern library, model-routing data, security bulletins, and agent capability benchmarks.
- friday digest of what shipped
- routing: which model is winning which task
- cancel anytime; local memory stays
C · hybrid
advanced playbooks.the playbooks
architectural-consistency rules, tier-aware agent configurations, and advanced governance.
- architectural rules that travel
- tier-tuned dispatch profiles
- cancel keeps what is in your repo
the surface.
three things that ship every day.
not a screenshot gallery. graph viewer, master plan, and doctor are the operational surfaces users inspect.
~/0dai · graph @ project=repo/auth
deliberation becomes a product surface.
The prototype makes invisible agent work visible: memory load, routing, consensus, patch hold, QA proof, and outcome capture.
Load local decisions and promoted patterns before code changes.
Pick the model that wins this task class this week.
Surface blockers before they become a bad merge.
~/repo/auth · 0dai
idle
→ 0dai run "ship auth"
✓ Roadmap synced ................. 24 decisions loaded
✓ Swarm handoff .................. 3 agents · one task
✓ Review receipt ................. tests + notes ready
$
start local, no account.
a free account adds the rest.
0dai init --local writes a CLAUDE.md / AGENTS.md scaffold with no signup. A free, no-card account adds the full per-CLI config set, sync, and run.
~/repo $ live install
$ npm i -g @0dai-dev/cli
fetching @0dai-dev/cli@4.4.6
$ 0dai init --local
detecting agents: claude-code, codex, gemini, aider, opencode, qoder
wrote CLAUDE.md · AGENTS.md · ai/manifest/ — no account needed
$ 0dai activate free
free account (no card) · adds the full per-CLI config set + sync + run
Install the CLI globally from npm.
Write the scaffold with one command. No account needed.
Free, no card. Adds the full per-CLI set, sync, and run.
Hand a real goal to 0dai run "..." --now and read the receipt.
Ship one real change and capture the outcome.
the whole CLI fits
in five commands.
No long manual. These five cover install-to-first-task. Each is the same line you will find in the README.
create ai/ layer, auth, and MCP bootstrap
run one task locally now, get a scored receipt — no account
health check for credentials, drift, and env
maturity, swarm tasks, and session state
refresh ai/ after repo or server changes
free stays free.
upgrade when you want the network.
Cancel Pro and you keep what is already in your repo. The cloud layer goes dark. No bait, no decay of local code.
Free
$0free account, no card. ships in every project.
- your repo stops paying re-explanation tax
- safety rules + master plan in one command
- 49 core MCP tools
- 6 terminal agent CLIs plus Cursor MCP bridge, one config
// the moat lives here
Pro
$15per month · per developer
- your project compounds across sessions
- 3-vendor deliberation before risky work
- routing that learns which model wins
- All 113 registered MCP tools
// patterns travel
Team
$49per month · per seat
- patterns travel across your repos
- shared memory graph for teams
- architectural rules enforced as agents code
- private bulletins for org-only patterns
Cost calculator
Estimate a monthly bill from PR volume, agent count, token budget, and runner choice.
Monthly estimate
$50
- PR processing
- $20
- Token spend
- $0.10
- Agent overhead
- $10
- Runner pool
- $20
Cursor-style seats$40
Devin-style flat plan$500
Estimates use transparent unit assumptions; final invoices follow authenticated account metering.
// pay withstripe · cardwallet payTONinvoice (team+)
sharp answers beat vague promises.
Is this just another CLAUDE.md generator?
No. Native configs are the entry point. The value is the living memory, routing, outcomes, and gates around the agents.
Does source code leave the machine?
The free local workflow keeps source local. Product metadata, manifests, and promoted patterns are the cloud boundary.
Why show pricing on the landing page?
The concept is explicit: free stays free, Pro is for the network, Team is for shared operational memory.