Platform Services (BaaS)
kis.ai isn’t only a CLI, it’s a Backend-as-a-Service (BaaS) platform. The
backend is a set of independent service binaries (the *.svc files installed by
kvm into ~/.kisai/bin/). Each has an internal codename, a
focused job, and its own command-line interface for running and bootstrapping it.
You operate them three ways: kvm installs/updates them,
kis info
inspects them. Several kis commands are clients of these services, e.g.
kis config/kis sd talk to infinity-config, and the
gateway provider talks to the AI Gateway.
The service map
Section titled “The service map”kvm install all (or kis-platform / baas-full) lays down ~35 service binaries.
They’re grouped below by function. Many carry an internal codename. Every service
also has a version subcommand (omitted from the tables for brevity).
Edge & gateway
| Binary | Role | Run/CLI |
|---|---|---|
gateway.svc | API Gateway, TLS-terminating, multi-tenant edge proxy | config |
ai-gateway.svc | AI gateway, one endpoint fronting many model providers (the gateway provider) | |
harbor.svc | Distribution server | keygen, serve |
AI & knowledge
| Binary | Role | Run/CLI |
|---|---|---|
ai-flow.svc | AI flow orchestrator | agent, local |
ai-bot.svc | Chatbot substrate: channel adapters, staged NLU cascade, response orchestration, sandboxed bot scripts | |
script.svc | Sandboxed multi-language script execution (the engine behind script/solve) | |
search.svc / know.svc | Lexical + semantic search |
Data & storage
| Binary | Role | Run/CLI |
|---|---|---|
data.svc | Data services / data-lifecycle layer | config, data |
datapipe.svc | Data-pipeline orchestrator | agent, local |
meta.svc | Component data serving (embed products into a binary) | config, embed, embed-info, serve |
ingest.svc | Capture traces & logs → logstore | generate |
backup.svc | Backup/restore, scheduling, and backup validation | |
content.svc | Content/media service (needs libvips) |
Workflow, jobs & rules
| Binary | Role | Run/CLI |
|---|---|---|
automate.svc | Distributed workflow orchestrator (coordinates vulcan agents) | agent, build, cli, local |
workflow.svc | Workflow orchestrator | agent, local |
jobs.svc | Environments, products, pipelines & job execution | agent, execute |
rules.svc | Business-rules execution engine |
Identity, security & secrets
| Binary | Role | Run/CLI |
|---|---|---|
iam.svc | Identity & access management (see IAM) | config, data, seed, tenant |
vault.svc | Certificate / secret services (see Vault) | cert, config, server |
vault-agent | Per-node Vault daemon | doctor, init, install, run, status, uninstall, upgrade |
audit.svc | Accepts and stores audit records from kis.ai services | generate |
argus | Runs profile-configured security scanners over source trees | consolidate, enrich, … |
Platform, config & ops
| Binary | Role | Run/CLI |
|---|---|---|
config.svc | Hive configuration + service discovery server (the config/sd backend) | config, explain, import-legacy, migrate, server |
supervisor.svc | Keeps a declared set of services in their desired state | daemon, install-daemon, install-service, restart, server, start, … |
forge.svc | Environments/products/pipelines management | generate |
root.svc | Environments/products/pipelines management | generate |
integrate.svc | Integrations management | generate |
code.svc | App generation from Meta templates (the code backend) | |
status.svc | Status APIs | generate |
Observability & comms
| Binary | Role | Run/CLI |
|---|---|---|
observe.svc | Query logs and traces | |
telemetry.svc | Telemetry APIs | generate |
notify.svc | Notifications across email, SMS, push, WhatsApp, in-app, webhook | generate |
rtc.svc | Chat / channels / teams |
Dev & test
| Binary | Role | Run/CLI |
|---|---|---|
test.svc | Test execution & automation | agent, api, load, orchestrator, test, tests |
bff.svc | Backend-for-frontend | generate |
scraper.svc | Scraping service |
Also installed: kis.ai, a monolithic all-in-one build with every CLI
command in one binary; llama-server, a bundled llama.cpp server for local model
serving.
Common patterns across services
Section titled “Common patterns across services”- Run the service,
server(infinity-config),serve(meta), or a mode likeagent/local/execute(AI Flow, Jobs). Workers typically run inagentmode; orchestrators run inlocal/execute/servermode. - Scaffold config,
generateprints or writes a curated example configuration forconfig.svc,data.svc, andmeta.svc. - Migrate / import,
migrateapplies a major-version schema migration andimport-legacyconverts legacy clusterstore/customerstore trees to the new hive layout on infinity-config. version, every service binary has aversionsubcommand.
# Examples (each binary lives in ~/.kisai/bin and is run directly)config.svc server # run the infinity-config serverconfig.svc generate # print a curated infinity-config examplejobs.svc execute # run Jobs in execute modejobs.svc agent # run a job workerdata.svc generate # print a Data API bootstrap examplemeta.svc serve # serve local folders / git repos as componentsai-gateway.svc --help # AI Gateway optionsInstalling the BaaS stack
Section titled “Installing the BaaS stack”Services come as kvm bundles:
kvm bundles # list bundleskvm install baas-core # essential BaaS serviceskvm install baas-full # all BaaS serviceskvm install ai-full # all AI serviceskvm install kis-platform # BaaS + AI + dev tools (everything)kvm install iam.svc # a single service (here: IAM)Then inspect what’s running with
kis info. Each service’s own --help documents its config files and
run modes.