Opensolr Changelog

Every integration can now be tried without an account. The LangChain, LlamaIndex, Haystack, MCP and Laravel Scout packages all ship with the credentials of a public demo account, so the quickstart in each README runs on a fresh machine with nothing to sign up for and nothing to configure. An index preloaded with 300 news articles is there to search immediately, and you can create your own indexes on the account and push documents into them. Everything created on it is deleted after 3 days, the account is shared with everyone using it, and the limits are small and per index — 200 MB of bandwidth and 50 MB of disk — so it covers a demo and a proof of concept rather than an application.

The LangChain, LlamaIndex, Haystack, MCP and Laravel Scout packages are out at 0.3.0, and every one of them is now covered end to end by a live test suite that runs against the real API. Three things that suite found are fixed in this release: a mistyped hybrid search mode returned a silently different result set instead of an error, argument checks ran after the embedding call rather than before it, and a dropped connection surfaced as a raw network error instead of being retried.

api.opensolr.com now serves only the AI and Data Ingestion endpoints. Embeddings, hybrid search, ai_summary and the ingestion queue answer there; every management endpoint (indexes, configuration, crawler, backups, security) is served exclusively by opensolr.com and answers 403 on the AI host. No client change is needed — all official integrations already use the right host for each call.

remove_http_auth works again. The endpoint that removes HTTP Basic Auth from an index had been failing on every call; it now removes the credentials on the index and on every cluster replica, exactly like update_http_auth sets them. Also fixed: the short https://api.opensolr.com/api/METHOD alias answered 404, and api_query_suggestions returned an error when called on opensolr.com.

The API Reference page now tells the truth about errors and hosts. It explains which endpoints live on opensolr.com (management) and which on api.opensolr.com (AI & Data Ingestion), lists the four HMAC-signed endpoints, and documents the real error responses (HTTP 200 with "status": false for application errors, 403 for authentication, 429 with Retry-After). Two commands that referenced non-existent endpoints were removed from the Solr Migration guide.

The whole REST API is now documented — 14 new endpoint articles. Every user-facing endpoint has its own page with parameters, real responses and cURL/PHP/Python examples: Hybrid Search (embed_and_search), vector_regions, get_account_summary, elevation_toggle, add_crawl_url_signed, delete_crawler_url, crawl_schedule_status, move_done_to_todo, tail_crawl_log, log tails, live streams, restore_backup, remove_http_auth and delete_environment. The Data Ingestion Queue page now lists every queue endpoint, and endpoint aliases (create_core, delete_index, reset_core, server_country) are noted on their articles.

All AI integrations updated. LangChain, LlamaIndex, Haystack, the MCP server and Laravel Scout all ship the new answer behaviour. Update with pip install -U or composer update.

The ai_summary API is now fully deterministic: your instruction is the prompt. Whatever you send in instruction is exactly what the model reads, byte for byte. Nothing reorders it, relabels it, appends to it or shortens it, so a prompt you tuned behaves the same on every call. Retrieval stays yours: run embed_and_search, take your top results and place them in the prompt where you want them. The query and context parameters still work for existing integrations. Prompts up to 100,000 characters are accepted.

Every AI integration ships the improved answer behaviour. LangChain, LlamaIndex, Haystack, the MCP server and Laravel Scout all now send a grounded answer instruction by default and drop weakly matching documents before generating, so RAG answers are direct and stay on topic. Update with pip install -U or composer update.

embed_and_search now matches your hosted search page exactly. The one-shot hybrid search endpoint runs the platform’s full tuned pipeline: your index’s saved Search Tuning applies automatically, natural-language questions are processed the same way as on your search page (stopwords cleaned for keyword matching, while the full phrase goes to the semantic embedder), and every tuning knob (fw_*, mm, search_mode, vector_topk, quality_boost, and more) accepts per-request overrides. Results returned via the API are now identical to your search page. Full parameter reference in the AI-Hints API guide.

Your Search Tuning now applies everywhere — including all AI integrations. Everything you configure in Index Settings → Search Tuning (semantic↔lexical balance, field weights, minimum match, search mode, vector candidate pool, content quality boost) now applies automatically to every search and every RAG answer made through the LangChain, LlamaIndex, Haystack, Laravel Scout, and MCP integrations — the same tuned pipeline as your hosted search page. Each knob can also be overridden per call from code via the new tuning parameter. Packages updated: LangChain, LlamaIndex, Haystack (v0.2.5), Laravel Scout (v0.2.5), MCP (v0.2.6).

Grounded RAG answers in every AI integration. LangChain, MCP, LlamaIndex, Haystack, and Laravel Scout now include a one-call ai_answer: hybrid (BM25 + kNN) retrieval picks your top documents, and their content becomes the LLM context — the same pipeline that powers the AI answers on Opensolr hosted search pages. Configurable: how many documents and how many words feed the LLM, plus an optional custom instruction (answer language, citation style, extraction tasks). The AI-Hints API guide now documents this two-step retrieve-then-summarize pattern with full cURL, PHP, and Python examples.

All AI integrations upgraded to v0.2.0. Writes in every integration — LangChain, MCP, LlamaIndex, Haystack, Laravel Scout — now go through the Data Ingestion API: documents are queued and enriched server-side (embeddings, sentiment, language, and all derived fields — identical to the Web Crawler), with progress visible in your Control Panel. Every integration also gains a lexical-only mode: pure keyword search with zero AI quota that works on any Opensolr index, including non-vector ones. Plus delete-by-query support and server-side PDF/DOCX text extraction via rtf:true.

Laravel Scout driver released. Add semantic + keyword search to any Laravel app: composer require opensolr/laravel-scout-opensolr, set SCOUT_DRIVER=opensolr, and Model::search() runs hybrid BM25 + kNN on a managed Solr index — with embeddings computed server-side. One index serves all your models. See the source on GitHub.

Haystack integration released. Opensolr now plugs into Haystack pipelines as a DocumentStore with a hybrid retriever: pip install opensolr-haystack. No embedder components needed — documents and queries are embedded server-side on our GPU infrastructure, and retrieval fuses BM25 + kNN scores natively. See the PyPI package and the source on GitHub.

LlamaIndex integration released. Opensolr is now a native LlamaIndex vector store: pip install llama-index-opensolr gives you OpensolrVectorStore and OpensolrEmbedding — server-side GPU embeddings, hybrid BM25 + kNN query mode, and standard metadata filters, on a managed Solr 9 index. See the PyPI package and the source on GitHub.

Opensolr is now in the official MCP Registry. Any AI agent that speaks the Model Context Protocol — Claude Desktop, Claude Code, Cursor, Windsurf and more — can now use your Opensolr indexes as native tools: hybrid BM25 + kNN search, document indexing with server-side embeddings, and grounded RAG answers. One config entry with uvx opensolr-mcp and your API key. See the registry listing, the PyPI package, and the source on GitHub.

New API endpoint: vector_regions. Returns the live list of vector-enabled environments (Solr 9.x with dense-vector schema and the hybrid query parser) — currently Chicago, Germany, and Finland. Tools like the LangChain package use it to discover valid locations automatically, so new regions work without any client updates. Need a dedicated vector-enabled environment in another region? We deploy them on request.

Opensolr is now a native LangChain vector store. Install with pip install langchain-opensolr and every LangChain chain, agent, and RAG tutorial runs on a managed Solr 9 index — with embeddings computed server-side on our GPU infrastructure (multilingual, 1024 dimensions) and true hybrid BM25 + kNN search with a tunable semantic-lexical balance. No embedding model to configure, no third-party API keys. See the LangChain integration page, the PyPI package, and the source on GitHub.

AI and vector features (AI Hints, AI Reader, Vector Embeddings, Hybrid Search) are now gated per package. Indexes without vector access get a friendly upgrade prompt instead of a generic error. Lexical keyword search continues to work on every plan.

New is_query parameter on the embed and batch embed API endpoints. Set is_query=1 when embedding search queries to get optimized retrieval vectors.

The Data Ingestion API now returns a doc_ids array in every successful response, showing the auto-generated document ID (md5 of uri) for each document in your batch. Use these IDs to track, query, or update specific documents in your index.

URI is now mandatory for every document in the Data Ingestion API. The document ID is always generated as md5(uri), making the URI the single source of truth for document identity. Same URI = same document. Resubmitting a URI updates the existing document. Duplicate URIs in pending jobs are automatically rejected to prevent accidental double-indexing.

JSON file upload for the Data Ingestion API. Instead of embedding your documents in the request body, save your full payload as a .json file and upload it via the payload_file field. The file can contain everything — email, api_key, core_name, and documents — or just the documents array with auth as separate form fields. Ideal for large batches generated by CMS exports or data pipelines.

Document extraction in the Data Ingestion API — add rtf:true to any document and point uri at a PDF, Word, Excel, PowerPoint, or OpenDocument file. Text is extracted automatically and indexed with full enrichment. Mix regular and RTF documents in the same batch.

Data Ingestion API — push documents directly into your Opensolr index via POST. Submit up to 50 docs per batch with automatic vector embeddings, sentiment analysis, language detection, and all derived search fields. Works alongside the Web Crawler to push content it cannot reach: gated pages, CMS exports, product feeds. Same doc ID updates existing records.

Global API rate limits are now enforced: 30 requests per minute and 500 per hour across all authenticated API endpoints. This applies to every API call including search, AI summaries, embeddings, and batch operations. If you need higher throughput for your integration, contact us and we will set custom limits for your account.

API rate limiting is now active. Every authenticated API request is tracked per user and per IP, with configurable limits per minute and per hour. If you exceed the threshold you will receive a 429 response with a Retry-After header — just back off and retry. Admin accounts are exempt. Individual users or IP addresses can also be given custom limits or blocked entirely. Monthly request counters are stored persistently so you can review your API usage over time.