20 Weights & Biases alternatives
Weights & Biases: The standard experiment tracker for machine learning, now with tracing and evaluation for LLM applications. 20 apps do similar work, most sharing Evaluation, Observability and Python, ranked by how much they have in common with Weights & Biases.
-
Open-source LLM observability — trace, evaluate, and improve AI applications with production data.
Also: Evaluation, Observability, Python
-
The framework and observability platform most teams use to build, debug, and ship LLM agents.
Also: Evaluation, Observability, Python
-
Warehouse-native product analytics that joins AI agent traces to user behaviour, so you can see which tool calls actually convert.
Also: Dashboards, Observability
-
Arize AI
AI engineering platform for tracing, evaluating and improving LLM apps and agents — with Phoenix, its open-source local-first counterpart, for teams that want to start without a vendor.
Also: Evaluation, Observability
-
Reads your production chat and voice agent conversations to surface the silent failures, frustration loops and policy breaches that offline evals never catch.
Also: Evaluation, Observability
-
Evaluation-first AI development — measure whether a prompt or model change actually improved anything.
Also: Evaluation, Observability
-
Prefactor Pty Ltd
Agent observability and evaluation that scores every production run for quality, drift and risk — and can pause or block a risky action before it executes.
Also: Evaluation, Observability
-
Agentic data platform that turns business questions into defensible, auditable narratives — with sources, assumptions and lineage attached.
Also: Dashboards
-
AI-native business intelligence: ask questions in plain English, get governed SQL, dashboards and daily insights across 750+ data sources.
Also: Dashboards
-
Teable AI, Inc.
AI spreadsheet-database on Postgres: describe what you need, and it builds the tables, automations, dashboards and internal apps around your data.
Also: Dashboards
-
Vectify AI
A vectorless RAG engine that builds a tree index of long PDFs and lets an LLM reason its way to the right section, with page-level citations.
Also: Python
-
The embedding database that starts as a Python import and grows into a distributed cloud service.
Also: Python
-
Run agent evals and private benchmarks in sandboxed cloud environments — compare Claude Code, Codex, Cursor and Copilot on the same real tasks.
Also: Evaluation
-
Abhishek Iyer
A local, keyboard-driven diff viewer built for the huge changes coding agents produce - and it remembers what you have already seen.
Also: Python
-
Microsoft
Microsoft's open-source optimizer that trains an agent's SKILL.md like a model parameter — epochs, learning-rate budgets and a held-out validation gate, with frozen weights.
Also: Evaluation
-
Open-source testing and evals platform for MCP servers — trace tool calls, debug OAuth, and catch cross-client regressions in CI.
Also: Evaluation
-
Tatva Labs Inc.
Continuous automated QA for voice and chat AI agents, from pre-launch simulation to live call monitoring.
Also: Evaluation
-
IBM (DataStax)
Open-source, MIT-licensed visual builder for AI agents, RAG apps and MCP servers, maintained by IBM's DataStax and extensible in Python.
Also: Python
-
HKUDS
Open-source, self-hosted personal AI agent you run yourself — browser UI, terminal or your chat apps, with memory, MCP tools and scheduled automations.
Also: Python
-
Lenz IO
Audit-grade fact-checking API for AI products — eight models across five stages, adversarial debate, and a full citation trail behind every verdict.
Also: Evaluation