DSPy
Stanford framework that compiles and optimises prompts instead of hand-tuning them.
why this verdict
Keep it — Anthropic has not replaced it
Anthropic's version overlaps, but does not finish the dev tools job, so this one is still worth keeping open.
- A named launch, not a vibe named, unlinked
- Frontier models becoming good enough that hand-written prompts rarely need optimising. Anthropic, August 5, 2025 — no announcement link recorded yet.
- How much of the job it covers not the job editorial call
- Parts of it. The job still needs the tool to get finished.
- Is there a free way to do it? yes
- 2 of 3 listed replacements have a usable free tier: TextGrad and Instructor.
- What the call is worth nothing to cancel
- No paid entry tier tracked, so there is no subscription to cancel.
- Threatened by
- Anthropic
- Since
- August 5, 2025
- List price
- free
- Per year
- —
The backstory
DSPy treats prompting as a compilation problem: you declare the signature of what you want, supply examples and a metric, and it searches for the prompt and few-shot demonstrations that maximise the score. The obvious objection is that better models need less prompt engineering. The counter is that they need more evaluation, and DSPy's optimisers are one of the few principled ways to improve a pipeline without a human guessing, which matters most when swapping models mid-project.
Escape hatches
Gradient-style optimisation over text pipelines
textgrad.com open_in_newEvaluation platform if you would rather tune manually
braintrust.dev open_in_newSimpler when you only need reliable structured output
python.useinstructor.com open_in_new2 of 3 replacements have a usable free tier.
