By the numbers

Data collected on 2026-04-30 from the master branch at commit 04e7a55f0b.

Size

The repository contains roughly 355,000 lines of Python across the workspace, split between source and tests.

Category	Lines	Files
Source (libs/**/*.py excluding tests)	~209,000	1,702
Tests (libs/**/tests/**/*.py)	~145,000	768
Total Python	~354,000	2,470

The repo holds 21 pyproject.toml files, each defining an independently versioned package with its own uv.lock.

xychart-beta horizontal
    title "Source files by package (top 12)"
    x-axis ["langchain-classic", "langchain-openai", "langchain-core", "langchain-anthropic", "langchain-huggingface", "langchain-fireworks", "langchain-mistralai", "langchain-groq", "langchain-ollama", "langchain-qdrant", "langchain-perplexity", "langchain (v1)"]
    y-axis "Files" 0 --> 1400
    bar [1322, 21, 183, 17, 13, 9, 7, 8, 7, 7, 11, 33]

Activity

The repo has 15,828 commits since the initial commit on 2022-10-24. 637 commits landed in the last 90 days — roughly 49 commits per week, with feat/, fix/, and chore/ types in roughly equal proportion thanks to dependabot and integration-test automation.

Window	Commits
Total since 2022-10-24	15,828
Last 90 days	637
Last 30 days	~200

Recent monthly commit counts (last 12 months):

Month	Commits
2025-05	198
2025-06	167
2025-07	288
2025-08	163
2025-09	248
2025-10	347
2025-11	141
2025-12	199
2026-01	181
2026-02	195
2026-03	233
2026-04	199

Commits in the last 90 days, by package

Path	Commits
libs/core/	136
libs/langchain_v1/	100
libs/partners/openai/	92
libs/partners/anthropic/	66
libs/langchain/ (classic)	62
libs/text-splitters/	52
libs/model-profiles/	44
libs/standard-tests/	37

langchain-core is the busiest area: every change there potentially affects every downstream package.

Bot-attributed commits

Roughly 35% of commits in the history have a noreply or bot author/committer signature (5,579 of 15,827). The bulk of these are dependabot and github-actions[bot] chores, so the figure is a lower bound on automation involvement and not a measure of AI-assisted human contributions.

Contributor diversity

Metric	Value
Unique commit authors	~3,572
Median commits per author	1 (long-tail of one-off contributions)

This is consistent with a popular open-source library — most contributors land a single PR (typically a small integration fix or doc tweak), while a small core team handles the bulk of the framework code.

Complexity

The largest source files in the workspace are concentrated in two areas: the runnables core and partner chat-model implementations.

File	Approx. lines
libs/core/langchain_core/runnables/base.py	~5,800
libs/partners/openai/langchain_openai/chat_models.py	~2,400
libs/partners/anthropic/langchain_anthropic/chat_models.py	~2,200
libs/core/langchain_core/messages/utils.py	~2,200
libs/langchain_v1/langchain/agents/factory.py	~1,875
libs/langchain_v1/langchain/agents/middleware/types.py	~1,700
libs/core/langchain_core/messages/content.py	~1,100
libs/langchain_v1/langchain/chat_models/base.py	~1,000
libs/core/langchain_core/language_models/chat_models.py	~2,500
libs/core/langchain_core/language_models/llms.py	~1,400

Test-to-source ratio

Across the workspace:

Package	Source files	Test files	Ratio
langchain-core	183	many (tests/{unit,integration,benchmark})	~1:1
langchain-classic	1,322	many	~0.5:1
langchain-openai	21	dozens of unit + cassette files	~3:1
langchain-anthropic	17	many	~3:1

Partner packages favor a high test-to-source ratio because every model release needs new behavioral tests.

Releases

The repository releases packages independently. Each package's tag namespace is <package>==<version>. The most recent tags include:

langchain==1.2.16
langchain-core==1.3.2
langchain-anthropic==1.4.2
langchain-openai==1.2.1
langchain-perplexity==1.2.0
langchain-fireworks==1.2.0
langchain-text-splitters==1.1.2
langchain-classic==1.0.4
langchain-tests==1.1.7
langchain-huggingface==1.2.2

Releases are triggered manually via the _release.yml workflow with working-directory and release-version inputs.

Dependencies

Top runtime dependencies declared in libs/core/pyproject.toml:

• langsmith>=0.3.45,<1.0.0
• pydantic>=2.7.4,<3.0.0
• tenacity!=8.4.0,>=8.1.0,<10.0.0
• jsonpatch>=1.33.0,<2.0.0
• PyYAML>=5.3.0,<7.0.0
• typing-extensions>=4.7.0,<5.0.0
• packaging>=23.2.0
• uuid-utils>=0.12.0,<1.0
• langchain-protocol>=0.0.10

libs/langchain_v1/pyproject.toml adds only:

• langchain-core>=1.3.2,<2.0.0
• langgraph>=1.1.10,<1.2.0
• pydantic>=2.7.4,<3.0.0

That's the entire required dependency set for langchain itself — partner packages opt in via optional extras.