Practical build guides for the BisenseAI platform—backend logic canvas, Weaver frontend studio, and deployable APIs. Each guide walks through real nodes, integrations, and production patterns.
Ingest PDFs and wikis on the BisenseFlow backend canvas, embed into Pinecone or Chroma, and expose a Weaver search UI—no custom RAG scripts. This guide covers semantic chunking, metadata RBAC, re-indexing, playground validation, and deploy as API or MCP.
Let users define Agent tools visually—HTTP nodes, JSON schemas, and Composio integrations—then attach them to agents without redeploying code. Covers validation, tenant isolation, sandboxing, Weaver test playground, and MCP packaging.
Chain a fast LLM rewrite on BisenseFlow before your main model, with a Weaver side-by-side prompt comparison UI. Style presets via Logic branches, cost control with small models, and deploy as POST /enhance API.
Combine HTTP nodes, Composio, LangChain tools, and custom Python on BisenseFlow—expose them to agents and MCP clients. Library of small tool workflows, schema design, retries, observability, and marketplace packaging.
Deploy BisenseFlow workflows as MCP servers so Claude and other clients discover tools and resources with standard manifests—not custom glue. Tool vs resource workflows, SSE vs stdio transport, auth, and Weaver admin toggles.
Use Agent nodes, loops, max-step guards, and Vector Store memory on BisenseFlow—plus Weaver approval gates for sensitive actions. Bounded autonomy, HITL interrupts, multi-agent handoffs, and production monitoring.
Model multi-stage pipelines with branches, parallel macro nodes, and checkpointed state—debug with time-travel on BisenseFlow. Enterprise DAGs: fan-out research, QA loops, HTTP checkpoints, and ops dashboards on Weaver.
Loop CSV keywords on BisenseFlow: LLM SEO copy, optional fal.ai heroes, Weaver React templates, deploy via Composio/GitHub. Programmatic SEO with uniqueness critic, schema validation, and throttled batch deploy.
Weaver block editor wired to BisenseFlow rewrite, summarize, and tone branches—with streaming LLM output and time-travel debug. Action-specific subgraphs, selection guards, rate limits, and markdown-preserving prompts.
Chain STT → LLM → ElevenLabs HTTP TTS on BisenseFlow with FFmpeg merge and Weaver record/playback UI. Sentence streaming, barge-in abort, dubbing mux, and tool-calling voice agents.
fal.ai and Replicate nodes for generation, prompt-enhancer LLM branch, S3 File Output, and Weaver gallery with async webhooks. Safety gate, inpainting workflow, CDN archival, and batch loops for catalog creatives.
Chain fal.ai/Replicate video, ElevenLabs TTS, FFmpeg transcode, and DaVinci Resolve scripting on BisenseFlow—from script JSON to platform-ready MP4. Covers scene loops, async webhooks, loudness normalization, 9:16/16:9 branches, Weaver progress UI, and production cost controls.
Generate React layouts on Weaver from LLM component JSON, validate with Playwright screenshots, and push to Figma via Composio. Covers design tokens, HTML/React import, visual QA LLM, and deploy preview URLs for stakeholder review.
Master the BisenseAI dual canvas: BisenseFlow for agents and integrations, Weaver for UI—shared I/O, time-travel debug, API and MCP deploy. Reference architecture for every product team shipping AI-native apps on one runtime.
Repo RAG on BisenseFlow, Tools Agent with safe patch tools, Composio GitHub PR, Weaver diff UI, and CI test nodes. White-label coding assistant you own—models, policies, and observability included.
Playwright trend research, LLM channel-specific copy, fal.ai images, HITL approval on Weaver, Composio publish to LinkedIn/X/Instagram. Brand-safe autonomous marketing with kill switch and engagement analytics loop.
Scheduled Playwright harvest, LLM summarize/tag, vector dedup, Weaver inspiration board with semantic search—feed your content pipeline.
Editorial pipeline on BisenseFlow: outline → draft → SEO critic loop → fal.ai hero → HITL on Weaver → Composio CMS publish. State object carries brief through stages with max two SEO revision loops.
Parallel Composio fetches (GA4, Meta, LinkedIn, Mailchimp), DataFrame merge, LLM executive narrative, Weaver KPI dashboards, Gmail digest.
Production RAG on BisenseAI: query router, multi-query expansion, hybrid retrieval, reranker, CRAG grader, SQL bypass, RBAC metadata, citation UI.
Batch-process CSV leads, SKUs, or keywords with BisenseFlow loops/macros—per-item state isolation, concurrency caps, merge CSV, error column.
Replace god-prompts with typed step interfaces on BisenseFlow—schema validation between LLM nodes, Python for math, Weaver step checklist UI.
