Scale AI coding without losing control.

    Govern every user, model, dollar, and data boundary from one enforcement layer, inside your cloud or ours.

    Talk to usBook a demo
    Kimchi deployment topology inside your VPC
    AkamaiWeWorkRokuCiscoYotpoSwisscom

    Your cloud. Your data. Your rules.

    Open models run inside your AWS, GCP, or Azure account. Prompts, completions, and code never leave your cloud.

    Other AI vendors

    Your code, their cloud.

    Closed APIs send every prompt, file, and secret to infrastructure you do not control and jurisdictions you may not be allowed to use.

    your code→their server→their model

    Kimchi Enterprise

    Your code, your perimeter.

    Every approved model runs inside your VPC. You decide what crosses the boundary; the default is nothing.

    your code→your VPC↺your model
    Read more about security

    Built for regulated teams.

    Identity, audit, and compliance controls sit on top of a VPC-native architecture, with proof for every policy.

    Identity

    Identity & access

    SSO, SAML, OIDC, group-based RBAC, and SCIM-ready provisioning tied to your identity provider.

    Audit

    Every request traceable

    Record every prompt, completion, and tool call. Ship logs to your SIEM and retain an immutable audit trail.

    Secrets

    Credentials stay isolated

    Agents request scoped credentials at runtime from your secrets manager and never see raw keys.

    Data

    No training on your data

    No customer prompts, code, or outputs are used for model training in serverless or self-hosted modes.

    Compliance

    Enterprise-ready posture

    SOC 2 Type II, ISO 27001, and PCI DSS AOC, with HIPAA-ready architecture and BAA on request.

    Providers

    Explicit external routing

    Use your own provider keys with clear data flows and policy approval for every external model request.

    Full security overview

    No rogue merges. No leaked secrets. No runaway spend.

    Every coding agent routes through one policy layer before a request reaches a model.

    Coding agents
    Kimchi Coding
    Cline
    Windsurf
    OpenCode

    Kimchi enforcement

    Policy and control layer

    BudgetingHard caps per team, developer, or API key with a controlled fallback at threshold.
    Usage metricsPer-request traces across model, tokens, latency, tags, and cost.
    Skills registryOne approved catalogue of tools, prompts, and MCP skills for every team.

    Kimchi web app

    Reports and insights

    Analytics reportRequests, token consumption, spend trend, and breakdown by team.
    Adoption reportSeat utilization, active users, and onboarding progress over time.
    Coding agent reportSessions, completions, acceptance rate, and estimated time saved.
    Cost & billingUsage by model, credit consumption, and billable spend this period.

    See exactly where your AI spend is going.

    Track spend by developer, team, model, and tag in real time. Forecast before scaling and enforce limits before the invoice arrives.

    Kimchi Enterprise cost insights dashboard

    Use the right model for every task.

    Route planning and execution by complexity, sensitivity, quality, and team policy. Sensitive work defaults to approved models in your VPC.

    Kimchi Auto

    Model orchestration

    Each request is classified before a model is selected.

    Complex reasoning

    Policy-approved planning models

    Architecture, debugging, and long-context analysis route to an approved planning model when policy allows.

    reasoninglong contextplanning

    Self-hosted execution

    Open models in your VPC

    Generation, refactors, tests, and embeddings run on your GPUs at a lower unit cost.

    GLM 5MiniMax M2.7Kimi K2.5

    One command. Keep your existing workflow.

    Kimchi setup detects Windsurf, Cline, OpenCode, and VS Code, then updates compatible endpoints automatically. Start serverless and move to self-hosted when compliance or cost demands it.

    Kimchi enterprise migration setup

    Connect the tools your engineers already use.

    Bring approved context, persistent memory, and spec-driven workflows into one governed coding platform.

    GitHub Enterprise

    PRs, diffs, comments, and scoped status checks.

    GitLab

    Merge requests and pipelines, including self-managed instances.

    Jira

    Read tickets, link commits, and create issues from findings.

    Confluence

    Specs, runbooks, and ADRs available during planning.

    Slack

    Notifications and channel triggers with team routing rules.

    Linear

    Issues, cycles, and projects connected bidirectionally.

    Postgres / MySQL

    Scoped queries for retrieval and analysis.

    Cloud storage

    S3, GCS, and Azure document and artifact storage.

    Okta · SAML / OIDC

    SSO and RBAC mapped to identity-provider groups.

    Vault / Secrets Manager

    Credential isolation without exposing raw secrets.

    Datadog · Splunk

    Send prompts, completions, and tool calls to your SIEM.

    Custom MCP

    Turn an approved HTTP endpoint into a typed tool.

    The questions enterprises ask.

    Bring governed AI coding inside your perimeter.

    We’ll model your break-even, review the architecture, and build a deployment plan around your requirements.

    Talk to usBook a demo