Nasiko Labs
Nasiko builds an open runtime/control plane for AI agents, coding harnesses, frameworks and tools. It gives enterprises visibility into agent cost and behavior while providing routing, orchestration, runtime policy, security, observability and governance from local development through production.
At a Glance
- Enterprise engineering, platform and AI teams
- Developers and organizations operating coding-agent fleets
- Companies deploying multi-agent systems and A2A workflows
- Banking, telecom, logistics, SaaS and cloud organizations
- +1 more
AI Tools by Nasiko Labs
(1)
Nasiko
Open Source AI Agent Runtime
Discussions
No discussions yet
Be the first to start a discussion about Nasiko Labs
Latest News
Nasiko is now available as a one-click managed offering on the DigitalOcean Marketplace.
Nasiko documented a major release wave covering coding-agent discovery/reporting, TokenOps, compression, HITL, custom providers, version rollback and SSO/SCIM UI.
Nasiko published how it built security into the platform before needing it, following its Secure Open Source Fund work.
Nasiko published how it strengthened agent and MCP security after its GitHub Secure Open Source Fund security work.
Products & Services
An open runtime and control plane for registering, deploying, orchestrating, routing, securing, observing and operating AI agents and workflows across frameworks and providers.
Cost and token operations: unified spend attribution across harnesses, frameworks, models, agents, sessions, developers and teams, with forecasting, pricing, dashboards and cost-reduction controls.
Workflow composition and execution for chaining multiple agents into repeatable, traceable flows.
A permission-filtered tool surface with connectors, OAuth, uploads and per-agent allow/ask/block controls for MCP servers and tools.
Market Position
Nasiko positions itself as an open, vendor-neutral runtime/control plane rather than another agent framework: it sits in the call path beneath existing harnesses and frameworks, works across model providers, and combines cost attribution, routing, orchestration, governance, MCP/tool controls and observability. This differentiates it from framework-specific platforms and from narrower agent tracing, LLM gateway or token-cost products.
Leadership
Founders
Rahul Todkar
Co-Founder and CEO. His public profile describes him as an AI, technology and product executive and cites prior roles at LinkedIn, PayPal, Intuit, Charles Schwab, eBay and Tripadvisor, where he was VP/Head of Data & AI.
Karan Bharadwaj
Co-Founder and Head of Product. His public profile says he has built companies and products across AI and emerging technology; public professional listings cite prior work with Project NANDA, OpenGrowth Ventures, Arithmic Labs and Mool.
Ramesh Raskar
Co-Founder and Chief Scientist. He is an Associate Director of the MIT Media Lab, an Associate Professor at MIT, and director of MIT Media Lab's Camera Culture group.
Executive Team
Rahul Todkar
Co-Founder & CEO
AI, technology and product executive with prior roles including LinkedIn, PayPal, Intuit, Charles Schwab, eBay and Tripadvisor.
Karan Bharadwaj
Co-Founder & Head of Product
Product and technology builder with experience founding/building AI and emerging-technology companies and products.
Founding Story
Nasiko was started around the view that AI agents were becoming easy to build but lacked the operational layer needed to run reliably in production. Its initial vision was an open, vendor-neutral runtime/control plane that could register, discover, route, secure, observe and govern agents without forcing teams to replace their existing harnesses, frameworks or model providers.
Business Model
Revenue Model
The managed platform uses a flat monthly subscription with included agent compute and pay-as-you-go charges when usage exceeds the included pool. Nasiko also offers open-source/self-hosted software and usage-based production services such as execution monitoring, routing, governance and security controls.
Pricing Tiers
Includes a pool of agent compute; additional scale is pay-as-you-go.
Target Markets
- Enterprise engineering, platform and AI teams
- Developers and organizations operating coding-agent fleets
- Companies deploying multi-agent systems and A2A workflows
- Banking, telecom, logistics, SaaS and cloud organizations
- Open-source developers and self-hosting teams
- Enterprise AI-agent operations from local development to production
- Discovering, reporting and routing developer coding agents without replacing their harnesses
- Controlling AI-agent access to tools, MCP servers, secrets and other agents
- Monitoring multi-agent workflows, sessions, traces, failures and approvals
- Forecasting, attributing and reducing model/token spend
- Deploying agents built with containers, LangChain, CrewAI or A2A-compatible frameworks
- The MIT Media Lab event page reports participation from senior leaders at Cognizant, PayPal, Walmart and Glean; it does not identify them as Nasiko customers.