Hi Product Hunt 👋

I'm Jason Young, and I built Akema.

Why it exists: I built Akema because looking at a Kafka topic had turned into a project. A plain question — what does this message look like, when did the bad ones start, is it one producer — meant a throwaway consumer, a dump into storage, or a console that security would not put near production. Each of those leaves something behind: a script, a copy of the events, or a consumer group that commits offsets and can move someone else's place in the log.

Live traffic is not a warehouse. Akema is for the hour when you need to see what the topic is doing — debug a producer, check a schema, or walk a room through the stream — without standing up a pipeline just to look.

What it is: a live message inspector for Apache Kafka. You point it at a cluster, watch messages arrive in a table, and write one-line CEL rules ("traps") that tag the bad ones as they land. From there you can chart any field in the browser with DuckDB, or put a board on the big screen.

What it deliberately doesn't do: it's read-only. It never produces messages, commits your offsets, changes topics or ACLs, or stores payloads. Messages live in memory, and there's no telemetry. It's one binary with no JVM and no database.

Being upfront about where it stands:
- This is 1.0. Linux and Docker are the main platforms; the macOS and Windows builds are still marked beta.
- It's built for looking at a topic, not running a cluster. There's no topic admin, no consumer-lag management, and no producing test messages.
- Pricing: every download includes a 14-day trial of everything with no sign-up. After that, replaying CSV and JSON files stays free, and live Kafka is $99 per node per year. A license file works air-gapped.

No cluster handy? The free demo suite (linked from the docs) spins up Kafka in Docker with realistic traffic, so you can try it in a few minutes.

The feedback I'm after:
1. Does read-only matter to you, or would you trade it for features? Would it get Akema through change control where you work?
2. What would you look for first? Bad payloads, a schema change, a traffic spike, a single customer's messages?
3. Where did setup trip you up? Auth (SASL, mTLS), Avro or Protobuf, MSK, Confluent Cloud, Event Hubs. I'd like to know exactly where it failed.
4. Is $99 per node per year fair? If not, what would make it an easy yes?

Looking forward to your feedback.

Hi Product Hunt 👋 I'm Jason Young, and I built Akema. Why it exists: I built Akema because looking at a Kafka topic had turned into a project. A plain question — what does this message look like, when did the bad ones start, is it one producer — meant a throwaway consumer, a dump into storage, or a console that security would not put near production. Each of those leaves something behind: a script, a copy of the events, or a consumer group that commits offsets and can move someone else's place in the log. Live traffic is not a warehouse. Akema is for the hour when you need to see what the topic is doing — debug a producer, check a schema, or walk a room through the stream — without standing up a pipeline just to look. What it is: a live message inspector for Apache Kafka. You point it at a cluster, watch messages arrive in a table, and write one-line CEL rules ("traps") that tag the bad ones as they land. From there you can chart any field in the browser with DuckDB, or put a board on the big screen. What it deliberately doesn't do: it's read-only. It never produces messages, commits your offsets, changes topics or ACLs, or stores payloads. Messages live in memory, and there's no telemetry. It's one binary with no JVM and no database. Being upfront about where it stands: - This is 1.0. Linux and Docker are the main platforms; the macOS and Windows builds are still marked beta. - It's built for looking at a topic, not running a cluster. There's no topic admin, no consumer-lag management, and no producing test messages. - Pricing: every download includes a 14-day trial of everything with no sign-up. After that, replaying CSV and JSON files stays free, and live Kafka is $99 per node per year. A license file works air-gapped. No cluster handy? The free demo suite (linked from the docs) spins up Kafka in Docker with realistic traffic, so you can try it in a few minutes. The feedback I'm after: 1. Does read-only matter to you, or would you trade it for features? Would it get Akema through change control where you work? 2. What would you look for first? Bad payloads, a schema change, a traffic spike, a single customer's messages? 3. Where did setup trip you up? Auth (SASL, mTLS), Avro or Protobuf, MSK, Confluent Cloud, Event Hubs. I'd like to know exactly where it failed. 4. Is $99 per node per year fair? If not, what would make it an easy yes? Looking forward to your feedback.

Akema

Read-only Kafka message inspection & analytics

Read-only Kafka message inspection & analytics