TNAC.ai now points to the Joinable platform. The platform's core positioning is to solve the problem that "enterprises know far more than AI can use": key institutional knowledge in enterprises is scattered across carriers such as PDFs, emails, spreadsheets, manuals, support tickets, and chat records. This content was originally created for humans rather than machines and is difficult for AI to use safely and reliably. Through its platform, Joinable transforms this raw, scattered data into clean, governed knowledge that can be used by agents, and continuously keeps it in sync with data sources.
In terms of main capabilities, the platform includes a component called Propagator, which is responsible for distilling key institutional knowledge into clean, governed, agent-facing knowledge, maintaining continuous synchronization with sources, and inheriting the permission settings of the sources, so that agents can act based on trusted knowledge. In terms of trust and governance, each Data Card inherits the permissions of its source, and every request from an agent is authorized according to these permissions and recorded in a tamper-proof audit log; agents will never see content that the people behind them are not authorized to view. The platform supports connecting to any agent framework via MCP or REST, and has features such as permission governance, continuous synchronization, source traceability, tamper-proof audit logs, scoped delivery, and verified and scored.
The platform also provides agent framework capabilities, allowing teams to build and own agents that can reason based on what the enterprise already knows, including an orchestration runtime that can launch instances to connect agents, knowledge, and systems. The official emphasis is: "We don't build your agents, we build the things that make them trustworthy." In addition, the platform provides solutions for specific industries, such as Joinable Runbooks in the security field, which can transform security manuals and SOPs into governed, agent-facing knowledge, and then build and own response agents based on the enterprise's own approved processes rather than the open internet, with a human-in-the-loop model by default. The platform also provides a free adversary intelligence Threat Map, mapped to MITRE ATT&CK. The official statement says it has supported more than 140,000 AI projects.
In terms of usage, the materials mention that any agent framework can be connected via MCP or REST, and orchestration runtime instances can be launched to connect agents, knowledge, and systems, but no specific installation commands or detailed operating steps are provided. Applicable scenarios include enterprise knowledge management, security incident response, IT operations, legal, human resources, finance, and other fields, helping teams build agents based on trusted knowledge. In terms of limitations, the materials do not provide information such as fees, pricing, or specific platform support details, and none is currently provided.
