Helping regulated enterprises build and ship production-level AI agents as fast as startups.
San Francisco, CA
Joined May 2024
- Tweets48
- Following7
- Followers146
- Likes62
We started Jinba because a lot of internal automation work still means stitching together scripts, cron jobs, webhook handlers, and API glue. LLMs made it easier to describe what you want, but getting from a prompt to something a team can actually run safely is still messy.
With Jinba, you can describe a workflow in plain English, then edit it visually or in YAML. It connects to APIs and tools, and you can deploy the result as an endpoint, webhook, Slack bot, or scheduled job.
Jinba (YC W26) retweeted
The deadline to apply for the YC Summer 2026 batch is coming up in five days!
Apply by May 4th: ycombinator.com/apply
Jinba (YC W26) retweeted
@a16z @Techweek_ is coming to Boston, and then shortly after to New York City. We're @jinbaflow (YC W26), building the AI layer for the world's largest banks, and we'd love to organize an event or two in both cities during the tech weeks.
if you're a VC or startup or bank in Boston or New York City and would like to co-organize these events with us, please reach out to me ✌
also, if you're a student or technology enthusiast and you are interested in helping us run the events, get to meet some of the brightest people in Boston and NYC, and get some cool merch, plus allowance for your help, please reach out to me too :)
its time to builddd
Every manual step in a Tier 1 bank workflow compounds risk over time. Errors accumulate. Decisions slow down. Compliance exposure grows.
Jinba automates the full workflow deterministically, so every step is auditable, repeatable, and owned by your ops team.
Investment document assessment at Tier 1 banks involves dozens of manual review steps, each a potential bottleneck.
Jinba automates the full workflow deterministically. Every decision is logged, auditable, and explainable to your compliance team and regulator.
Contract teams at Tier 1 banks still extract clauses manually. It's slow, error-prone, and nearly impossible to audit at scale.
Jinba automates the full extraction workflow deterministically, with a complete audit trail and on-premise deployment.
KYC, compliance, and document review workflows — described in plain English, deployed in days, audited end-to-end. Available SaaS, VPC, or fully air-gapped on your own hardware.
probabilistic AI in a compliance workflow is a liability.
if your model can't tell a regulator exactly why it made a decision, it shouldn't be making the decision.
Jinba runs deterministic logic at every step. full audit trail. explainable by design. b
the April 2026 Wolters Kluwer findings are a wake-up call.
Federal examiners aren't just asking if you use AI.
they're asking if you can govern it. most programs fail due to 5 factors:
1. No audit trail on AI decisions
2. workflows owned by IT, not compliance
3. logic that can't be explained to a regulator
4. no deterministic controls tools
5. built for speed, not accountability
most automation tools get disqualified at Tier 1 banks before the demo ends. on-premise deployment is a hard requirement, and most vendors can't meet it.
Jinba deploys on-premise in days. book a call with us at jinba.io
at most Tier 1 banks, AI tools get bought by the tech team and killed by ops and compliance.
not because the tech is bad. because ops leads can't see the logic, can't audit the decisions, can't own the process.
Jinba was built for them. book a call at jinba.io
Power Automate fails at Tier 1 banks for one core reason: it wasn't built for compliance-grade requirements.
No deterministic logic. No full audit trail. No on-premise deployment.
Jinba handles all three. Book a call with us at jinba.io
AML analysts at Tier 1 banks spend hours every day triaging alerts that turn out to be false positives.
Jinba automates that triage logic deterministically, with a full audit trail, so every decision is explainable and nothing slips through.
jinba.io