Elasticsearch
Elasticsearch is Elastic's distributed search and analytics engine, not a leftover 7.10/8.0-only vector note, and not only a China marketing URL. LICENSE.txt is a triple license (AGPLv3-only, SSPLv1, Elastic License 2.0) by default; x-pack is Elastic License only. elastic.co/elasticsearch sells Search, observability, vector database, and Jina AI models as part of Elastic. Cloud paths are Serverless, Hosted, and self-managed, each with a free trial / download. elastic.co/pricing/serverless-search prints as-low-as VCU units: ingest $0.14/VCU-hour, search $0.09, ML $0.07, storage $0.047/GB-month, egress $0.05/GB. Point the website at the global product page, not /cn/.
Compare Vespa if you wanted another search+ranking engine with Cloud units, Qdrant if you wanted a dedicated vector engine, and Pinecone if you wanted Free / $20 / $50 min / $500 min.
Key Features
- Search engine first, vectors included: Live page lists vector database as one workload beside logs, search apps, and threat protection. Do not freeze kNN as a 7.10-only add-on.
- License is not Apache-2.0: Quote the triple license / Elastic License. Do not call the default tree MIT or Apache.
- Cloud is trial + units: Serverless search pricing on the public page is "as low as" VCU hours. Hosted and self-managed stay on elastic.co/pricing.
- Website should not be the CN path: Leftover
elastic.co/cn/elasticsearch/is dropped.
Use Cases
- People who will self-manage the download and accept AGPL/SSPL/Elastic License terms.
- Teams that will start a Cloud trial and then read VCU units.
- People comparing 77,791 stars against Vespa 7,052 or Qdrant 34,018.
Limitation: "As low as" VCU rows are not a quote for every region or vector profile. Confirm the calculator.
Pricing
GitHub and elastic.co/pricing/serverless-search on 2026-08-18.
| Piece | Price | Notes from first-party pages |
|---|---|---|
| Self-managed | $0 to download | Triple license / Elastic License. 77,791 stars. You run it. |
| Cloud trial | Free trial | Serverless and Hosted CTAs. Confirm current terms. |
| Serverless ingest | As low as $0.14 / VCU-hour | Search as low as $0.09; ML $0.07. |
| Serverless storage / egress | As low as $0.047 /GB-mo ; $0.05 /GB | Vector profiles get 50 GB free on that page. |
| LLM add-on on that page | $4.50 / $21 per 1M tokens | Playground / Assistant row. Confirm before quoting. |
If a leftover blog still says the latest vector story is 8.0 NLP, treat this table as the source.
Getting Started
- Open elastic.co/elasticsearch, not the leftover
/cn/URL. - Confirm elastic.co/pricing and the serverless-search unit page.
- Confirm Qdrant if you wanted a dedicated vector Cloud Free node.
- Do not paste 7.10/8.0 as "the latest version."
Frequently Asked Questions
Is it free?
Self-managed source is downloadable under the published licenses. Cloud is a trial plus units.
Is it Apache-2.0?
Not as the default. GitHub LICENSE.txt is AGPL + SSPL + Elastic License, with x-pack Elastic-only.
Same as Pinecone?
No. Pinecone prints Free / $20 / $50 min / $500 min. Elasticsearch Cloud on the public page is VCU "as low as" rows.
Alternatives
- Vespa: Apache-2.0 search/serving engine; Cloud has a unit table.
- Qdrant: Dedicated OSS vector engine plus Cloud.
- Pinecone: Managed vector DB with public dollar seats.
Tips
- Quote 77,791 stars and the triple license from the GitHub check.
- Point website at elastic.co/elasticsearch.
- Confirm VCU "as low as" before you treat $0.14 as the whole bill.
Conclusion
Elasticsearch is a 77,791-star search engine with AGPL/SSPL/Elastic License terms and Cloud VCU units, not a leftover 8.0-only vector blurb. Start at elastic.co/elasticsearch, then decide whether Qdrant or Vespa already covers the engine you need.
Comments
No comments yet. Be the first to comment!
Related Tools
Related Insights

Grok Bot and Hermes Bot: one person finally gets a think tank and a secretariat
Grok Bot now ships with Cursor Pro+. Hermes Bot runs on a VPS. They are not smarter chat boxes. The think tank advises, the secretariat executes, and you still make the call.
Hook OpenCode Go into Codex on Windows. Do Not Open a Second Toolkit.
A ChatGPT-signed Codex desktop app still shows mostly GPT in the picker. On Windows, enable only OpenCode Go and the Grok, GLM, Kimi, DeepSeek, and MiniMax models you already pay for appear in the same selector. Keys stay local. Native GPT stays put.
What Locks Codex Is the Picker, Not the Models
You already pay for OpenCode Go, Grok, and Z.ai, but the Codex picker still shows mostly GPT. The community project codex-router does not teach another install ritual. It puts subscriptions you already bought back into the selector, keeps keys on the machine, and leaves native GPT alone.