Agents running in minutes,on your machine.
Download one binary and you get indexing, search, OCR, and a swarm of agents at work.
Zero infrastructure.
Works with OpenAI, Anthropic, Gemini, DeepSeek, Mistral, and Ollama (with Hugging Face models)
Connect your keys
Native BYOK. Plug in OpenAI, Anthropic, Gemini, DeepSeek, and Mistral, or run locally with Ollama (Hugging Face models).
Compatible models
Use the model that fits each task.
Plug in your API keys and swap models per agent. Route light tasks to cheap options and reserve top models for what matters.
Technical features
Architecture built for devs who ship real agents.
Direct IDE integration, hybrid search, communication with external agents, and a 100% transparent API.
Right in your code — no MCP
Your agents talk directly to any code in your IDE. No MCP, no bridge, no friction.
Native Hybrid Search
BM25, Vector, and Graph combined into a single retriever. Lossless context recall.
True parallelism
Every agent is an isolated instance. Multiple agents working in parallel, for real.
Collaborative swarm
Agents coordinated with shared memory and learning across the swarm.
Multi-format evolving memory
Autonomous learning in any format — .md, PDF, JSON, YAML, XML.
Local LoRA/qLoRA training
Your own SLM trained right in the platform from the knowledge your agents accumulate.
Interop with external agents
Native communication with OpenClaw, Hermes, Nemoclaw, and any other agent in the ecosystem.
Native WhatsApp, no API
Direct WhatsApp connection without the official API. Telegram and Slack on the same layer.
An OpenSource Language built by agents.
DSL that models perception, reasoning, memory, and execution of your agents as a DAG. Memories are variables, skills are the standard library, concurrency is swarm delegation.
1flow "Web Q&A" {2 input question: Untrusted34 node "search_web" -> skill web_search(query: question, count: 3)5 node "search_memory" -> skill memory_search(query: question, scope: "shared")6 node "search_graph" -> skill graph_search(query: question, depth: 2)7 node "synthesize" -> skill llm.chat(prompt: "Answer using:", context: [..])8 node "verify" -> skill verify(source: synthesize.result)9 node "speak" -> skill tts.speak(text: synthesize.result)1011 search_web -> synthesize12 search_memory -> synthesize13 search_graph -> synthesize14 synthesize -> verify -> speak1516 on "user_message" -> search_web17}Memories as variables
Every named value is persistent by default. Every variable is agent memory — no boilerplate.
Hybrid 3-layer memory
BM25 (keywords) + Vector (768d embeddings) + Graph (relational). Queried in parallel, fused automatically.
Compile-time Trust Types
Secret, Untrusted, Fact, Hallucinable, Control. The parser prevents credential leaks and unvalidated data.
Frequently asked
Questions that come up most.
A complete AI stack with native tools and integrations to build and orchestrate swarms of agents, locally.
Yes, and forever. There are usage and agent/flow creation limits, but you can bring your own API key and use it completely free.
Yes. You connect your preferred model with your own API key. Tacflow works with hundreds of models — fully flexible.
Yes. It was designed to be local-first, with a clear path to Local AI Mode when it makes sense to reduce dependence on tokens and the cloud. Agents still use external models (from the app or via your key) to learn and look things up on the web.
Compatible with Windows, Linux, and macOS. Just 8 GB of RAM. No GPU required.
Yes. As your agents learn, they train their own local AI model and need fewer calls to external models. The more you use it, the more independent it becomes — with data staying 100% local.
Zapier and Make automate linear tasks. They only do what they were programmed to. Tacflow is multi-agent orchestration and automation — agents can make different decisions and solve different problems across runs.
In Tacflow you build 100% independent agents that truly work in parallel. Native evolving memory, autonomous Skill creation, and dozens of built-in tools — Vision, Listen, document and site OCR, WhatsApp without the API, Telegram, proprietary search engine, autonomous learning from sites, documents, and context, among others.