21 tools tracked
Avg 7d gain: +13
AI Frameworks
Build AI applications with the libraries and frameworks developers actually reach for. From LLM orchestration to agent toolkits.
Heat scores updated daily from 30+ live sourcesEmbed this scoreboard →
Movers this week
1
stable
rising
rising
peak
rising
rising
rising
rising
rising
peak
rising
stable
peak
peak
peak
declining
stable
declining
stable
fading
Candle
Teams & Solo
+6
7d
68
Heat
2
Together AI
Teams & Solo
+42
7d
64
Heat
3
Pinecone
Teams & Solo
+16
7d
53
Heat
4
Replicate
Teams & Solo
+1
7d
53
Heat
5
LiteLLM
Teams & Solo
+18
7d
52
Heat
6
Haystack
Teams & Solo
+21
7d
52
Heat
7
Burn
Teams & Solo
+30
7d
51
Heat
8
Instructor
Teams & Solo
+9
7d
46
Heat
9
Semantic Kernel
Teams & Solo
+17
7d
44
Heat
10
LangChain
Teams & Solo
+5
7d
43
Heat
11
Axolotl
Teams & Solo
+13
7d
37
Heat
12
ChromaDB
Teams & Solo
+10
7d
36
Heat
13
Unsloth
Teams & Solo
+4
7d
33
Heat
14
AutoGen
Teams & Solo
+2
7d
33
Heat
15
LlamaIndex
Teams & Solo
+3
7d
31
Heat
16
Weights & Biases
Teams
26
Heat
17
AI Visibility Lab
Teams & Solo
20
Heat
18
Weaviate
Teams & Solo
17
Heat
19
Qdrant
Teams & Solo
11
Heat
20
RunInfra
Teams & Solo
7
Heat
21
DSPy
Teams & Solo
3
Heat
Popular comparisons
Frequently asked questions
What belongs in AI Frameworks?
These are the libraries and orchestration layers developers use to build AI applications — LLM routing, retrieval pipelines, and agent toolkits. LangChain, LiteLLM, and LlamaIndex are tracked here. They're building blocks you import into your own code, not finished products you log into.
How do I decide between frameworks?
Weigh the abstraction against your control needs. A high-level framework gets a prototype running fast but can hide the retries, prompts, and token spend you'll eventually need to tune; a thinner router keeps you closer to the raw model calls. Look at each project's release cadence and issue activity, since a framework that stops moving in a fast-shifting field becomes a liability faster than most tools.