Automatically extract every file shared across Microsoft Teams channels and team sites — then search across all indexed content through a built-in Streamlit search portal.
ChunkIQ discovers all teams in your tenant and indexes files across all channel types automatically.
Files posted or shared in public channels. Automatically maps the underlying SharePoint team site and crawls the Files tab library.
Private channels with dedicated SharePoint sites. Each is discovered and crawled separately with appropriate permission scopes.
Cross-tenant shared channels and their associated document libraries are enumerated and included in the extraction run.
Every Microsoft Team has a backing SharePoint site. ChunkIQ crawls all document libraries on these sites, not just the default Files tab.
No per-team configuration needed. Add ChunkIQ once and every current and future team is covered.
Enumerates every team in the tenant automatically. New teams added after initial setup are automatically included on the next run.
Goes beyond the default Files tab to discover custom document libraries, wiki content libraries, and any other SharePoint libraries on the team site.
Extracts and processes files embedded inside Word, Excel, and PowerPoint documents found in Teams, tracking lineage back to the parent file and channel.
Each chunk is tagged with team name, channel name, site URL, library, file path, and modification date — enabling fine-grained filtering in search results.
Content hash deduplication means only new or changed files are re-extracted. Large tenants with thousands of files process efficiently on every run.
Runs entirely within your Azure tenant. Uses managed identity or service principal with the minimum required permissions.
ChunkIQ Teams ships with a Streamlit-powered search portal. Hybrid BM25 + vector + semantic re-ranking returns the most relevant chunks from every team, channel, and document.
Combines BM25 keyword scoring with 1,536-dimensional HNSW vector search. Results are re-ranked with Azure AI Search semantic ranking via Reciprocal Rank Fusion.
Every result shows the originating team, channel, file name, and folder path. Click straight through to the source document in Microsoft Teams.
The Streamlit app connects directly to your Azure AI Search index using the same connection settings as the pipeline. No additional backend required.
Authenticates via Azure AD with Group.Read.All and Files.Read.All permissions to access all teams and their underlying SharePoint sites.
Enumerates all teams, resolves their SharePoint team sites, and lists every document library — standard, private, and shared channels included.
Dedicated extractors process each file format. Text is cleaned, split into semantic chunks, and enriched with Teams-specific provenance metadata.
Chunks are embedded with Azure OpenAI and pushed to Azure AI Search for hybrid BM25 + vector + semantic search across all Teams content.
One setup. Every team. Every channel. Every file — automatically extracted, indexed, and searchable through the built-in portal.