Show HN: Ktx – Open-source executable context layer for data agents
Recorded: May 28, 2026, 6 p.m.
| Original | Summarized |
GitHub - Kaelio/ktx: ktx is the context layer for analytics agents · GitHub Skip to content Navigation Menu Toggle navigation
Sign in
Appearance settings PlatformAI CODE CREATIONGitHub CopilotWrite better code with AIGitHub SparkBuild and deploy intelligent appsGitHub ModelsManage and compare promptsMCP RegistryNewIntegrate external toolsDEVELOPER WORKFLOWSActionsAutomate any workflowCodespacesInstant dev environmentsIssuesPlan and track workCode ReviewManage code changesAPPLICATION SECURITYGitHub Advanced SecurityFind and fix vulnerabilitiesCode securitySecure your code as you buildSecret protectionStop leaks before they startEXPLOREWhy GitHubDocumentationBlogChangelogMarketplaceView all featuresSolutionsBY COMPANY SIZEEnterprisesSmall and medium teamsStartupsNonprofitsBY USE CASEApp ModernizationDevSecOpsDevOpsCI/CDView all use casesBY INDUSTRYHealthcareFinancial servicesManufacturingGovernmentView all industriesView all solutionsResourcesEXPLORE BY TOPICAISoftware DevelopmentDevOpsSecurityView all topicsEXPLORE BY TYPECustomer storiesEvents & webinarsEbooks & reportsBusiness insightsGitHub SkillsSUPPORT & SERVICESDocumentationCustomer supportCommunity forumTrust centerPartnersView all resourcesOpen SourceCOMMUNITYGitHub SponsorsFund open source developersPROGRAMSSecurity LabMaintainer CommunityAcceleratorGitHub StarsArchive ProgramREPOSITORIESTopicsTrendingCollectionsEnterpriseENTERPRISE SOLUTIONSEnterprise platformAI-powered developer platformAVAILABLE ADD-ONSGitHub Advanced SecurityEnterprise-grade security featuresCopilot for BusinessEnterprise-grade AI featuresPremium SupportEnterprise-grade 24/7 supportPricing Search or jump to... Search code, repositories, users, issues, pull requests...
Search Clear
Search syntax tips Provide feedback Include my email address so I can be contacted Cancel Submit feedback Saved searches
Name Query To see all available qualifiers, see our documentation. Cancel Create saved search Sign in Sign up
Appearance settings Resetting focus You signed in with another tab or window. Reload to refresh your session. Dismiss alert Kaelio ktx Public
Notifications
Fork
Star Code Issues Pull requests Actions Security and quality Insights
Additional navigation options
Code Issues Pull requests Actions Security and quality Insights
mainBranchesTagsGo to fileCodeOpen more actions menuFolders and filesNameNameLast commit messageLast commit dateLatest commit History405 Commits405 Commits.github.github assetsassets docs-sitedocs-site docsdocs examplesexamples packages/clipackages/cli pythonpython scriptsscripts skills/ktxskills/ktx websitewebsite .gitignore.gitignore .pre-commit-config.yaml.pre-commit-config.yaml .releaserc.cjs.releaserc.cjs AGENTS.mdAGENTS.md CLAUDE.mdCLAUDE.md CONTRIBUTING.mdCONTRIBUTING.md GEMINI.mdGEMINI.md LICENSELICENSE README.mdREADME.md SECURITY.mdSECURITY.md biome.jsonbiome.json codecov.ymlcodecov.yml conductor.jsonconductor.json knip.jsonknip.json package.jsonpackage.json pnpm-lock.yamlpnpm-lock.yaml pnpm-workspace.yamlpnpm-workspace.yaml pyproject.tomlpyproject.toml release-policy.jsonrelease-policy.json skills.sh.jsonskills.sh.json tsconfig.base.jsontsconfig.base.json uv.lockuv.lock View all filesRepository files navigationREADMEContributingApache-2.0 licenseSecurity The context layer for data agents Quickstart · ktx is a self-improving context layer that teaches agents how to query your Why ktx Learns from company knowledge. Ingests wiki content, organizes it, How ktx compares General-purpose agent Builds warehouse context automatically Detects joinable columns + resolves fan/chasm traps Approved, reusable metric definitions Absorbs wiki / Notion / team knowledge Flags contradictions across sources Ships CLI + MCP for agent execution Read-only by design Who is ktx for Want agents like Claude Code, Codex, Cursor, or OpenCode to query your Skip ktx if you: You don't have a SQL warehouse - ktx sits on top of one Works with PostgreSQL, Snowflake, BigQuery, ClickHouse, MySQL, SQL Server, and TipAlready using an agent? Ask Claude Code, Codex, Cursor, or OpenCode from ImportantIf ktx status prints ktx mcp start --project-dir ..., run it before First commands Command ktx setup ktx status ktx ingest ktx sl "revenue" ktx wiki "refund policy" ktx mcp start See the CLI Reference Commit ktx.yaml, semantic-layer/, and wiki/. Keep .ktx/ local. Does ktx send my schema or query results to a hosted service? Docs Quickstart Community Slack — ask questions, share what you're building, and chat with maintainers. Development Path packages/cli packages/cli/src/context packages/cli/src/llm packages/cli/src/connectors python/ktx-sl python/ktx-daemon Local development CLI: About ktx is the context layer for analytics agents docs.kaelio.com/ktx Resources Readme Apache-2.0 license Contributing Contributing Security policy Security policy Uh oh! There was an error while loading. Please reload this page. Activity Custom properties 185 0 7 Report repository Releases v0.7.0 Latest Packages
Uh oh! There was an error while loading. Please reload this page. Uh oh! There was an error while loading. Please reload this page. Contributors Uh oh! There was an error while loading. Please reload this page. Languages TypeScript Python JavaScript MDX CSS Shell LookML
Footer © 2026 GitHub, Inc. Footer navigation Terms Privacy Security Status Community Docs Contact Manage cookies Do not share my personal information You can’t perform that action at this time. |
The ktx project represents a context layer designed specifically for analytics agents, serving to teach these agents how to accurately query data warehouses by integrating approved metric definitions, joinable columns, and relevant business knowledge that the system automatically builds and maintains. This layer addresses the inherent difficulties general-purpose agents face when performing data-related tasks, as they tend to re-explore data warehouses on every query, invent their own metric logic, and generate results that deviate from established definitions. Unlike traditional semantic layers, which require constant manual upkeep and fail to absorb broader company knowledge, ktx automates these processes. The core functionality of ktx involves several integrated mechanisms that automate the contextual understanding of the data environment. First, ktx learns from company knowledge by ingesting content from sources like wikis and Notion, where it organizes information, removes redundancies, and flags any contradictions for subsequent human review. Second, it maps the entire data stack by sampling tables, capturing metadata and usage patterns, detecting joinable columns, and annotating data sources to ensure agents write more effective queries. Third, ktx constructs a semantic layer by combining raw tables and high-level metrics through an internal join graph that automatically resolves potential difficulties, such as chasm and fan traps, enabling agents to fetch required metrics declaratively rather than requiring them to reconstruct canonical SQL repeatedly. Finally, ktx serves these contextualized insights to agents by exposing Command Line Interface and MCP tools, facilitating combined full-text and semantic search across both the internalized knowledge sources and the semantic-layer entities. In comparison to other approaches, ktx offers a distinct advantage by automating the contextualization process. It automatically builds warehouse context, manages approved, reusable metric definitions, and absorbs scattered corporate knowledge from various sources. Moreover, ktx actively flags contradictions discovered across these disparate sources, a feature lacking in passive semantic layers. The system also provides direct capability for agent execution by shipping CLI and MCP functionality. ktx is designed for organizations that utilize data warehouses supporting various SQL dialects, including PostgreSQL, Snowflake, BigQuery, ClickHouse, MySQL, SQL Server, and SQLite, and integrates seamlessly with tools like dbt, MetricFlow, LookML, Looker, Metabase, and Notion. Its primary utility is for users who want agents, such as those powered by Claude Code, Codex, Cursor, or OpenCode, to interact with their data warehouses using strictly approved metric definitions, thereby ensuring agents reuse canonical SQL instead of inventing new logic for every prompt. Initial setup of ktx involves running specific commands to configure the local environment, establish connections to data sources, build the necessary context, and install agent integrations. The system utilizes a structured project layout, including files for project configuration, semantic layer definitions, wiki documentation, and raw source artifacts, all of which organize the contextual information locally. The system operates entirely locally, ensuring data security because the connections are read-only, and no data is written to the underlying database; the local MCP daemon runs on demand when an agent client requires it. This architecture ensures that the ktx context remains a localized, secure asset while providing a comprehensive, searchable surface for agents to query complex analytical needs. |