tecweave unifies documentation from distributed teams and components into a single, intelligent knowledge graph — ingesting any format, processing with LLMs to find relations and contradictions, and validating that your docs actually match your software.
tecweave is built around a composable pipeline of independent services. Each service handles one concern perfectly — and they compose into a unified knowledge platform.
All source formats are normalized into a common internal schema, decoupling downstream processing from source specifics.
Discovers semantic connections between topics across components. Creates a linked knowledge graph.
Surfaces conflicting information across your component docs and generates a contradiction report.
Converts the processed knowledge into multiple output formats for downstream consumers.
Actively calls the software described in your documentation and cross-checks actual behavior against documented behavior — flagging every discrepancy.
Six purpose-built services work together across the tecweave pipeline — independently deployable, composable, and extensible.
Accepts documentation in any format — Markdown, OpenAPI, man pages, HTML, PDF, Confluence exports, and more. All content is normalized into a shared internal schema, decoupling your sources from your processing pipeline.
Large language models analyze the standardized knowledge to extract semantics, identify topics, summarize sections, and prepare rich metadata for downstream services.
Discovers semantic connections between documentation topics across teams and components. Related concepts are linked into a navigable knowledge graph, making cross-team knowledge discoverable.
Compares statements across all ingested documentation to surface conflicting information. Every detected contradiction is documented in a structured report with source references.
Export processed knowledge into the format your downstream systems need. Generators transform the standardized data into multiple output schemas simultaneously.
The validation service reads CLI and API documentation, executes the described commands and API calls against the real software, and reports every discrepancy between documented and actual behavior.
The tecweave validation service closes the gap between what your documentation says and what your software actually does. It parses CLI help text and API specification documents, then actively invokes the described commands and endpoints against the real running software.
Every discrepancy — wrong flags, missing parameters, changed response schemas — is captured in a structured discrepancy report so your teams can act immediately.
Reads CLI help output and OpenAPI / API specs from your ingest pipeline.
Runs the actual CLI commands or sends HTTP requests to the API exactly as documented.
Diffs actual vs. documented behavior and generates a detailed discrepancy report.
mycli --version Returns expected semver string
mycli export --format json Flag --pretty documented but not recognized
GET /api/v1/items Response schema matches spec
POST /api/v1/items Field metadata.tags absent in response
DELETE /api/v1/items/:id Returns 204 as documented
Connect your distributed teams and components into a unified, intelligent knowledge hub. Start with tecweave today.