Regular search mode
Fast facts, thin context
Good for quick lookups, but broad questions can collapse into a tidy paragraph that hides uncertainty.
Powered by Claude Fable
Meet the Fable model. It researches before it writes.
Fable Research brings Claude's Fable model into Felo for questions that deserve more than a quick summary. It reads source material, tests competing claims, and turns the evidence into a report you can use.
Open Fable Research Claude Fable uses credits quickly.
The point
The special part is not a new interface trick. It is the model behavior: deeper reading, more skeptical reasoning, and cleaner synthesis.
Fable workbench
Inside Fable's research loop
FABLE MODELRESEARCH
1
Reads the actual material
Fable is useful when the important part is buried in methods, definitions, caveats, or a long source.
2
Keeps a research thread alive
It uses one finding to shape the next step instead of treating every source as a separate summary.
3
Pushes back on weak claims
It marks missing context, fragile assumptions, and places where sources point in different directions.
4
Writes toward a usable report
It aims to produce a brief with findings, limits, implications, and next questions. Not just notes.
Model
Claude Fable
Designed for
Hard research questions
Best output
Evidence-led reports
Why it matters
The advantage is model behavior, not more links
Most research tools show you more results. Fable Research is different because the model spends more of the answer on interpretation, judgment, and synthesis.
Fable mode
Evidence plus judgment
Better for questions where you need to know what the sources really support, what they do not, and what to do next.
That is why the page leads with the model, not a feature list.
The better research result comes from Fable's behavior: careful reading, multi-hop reasoning, skepticism about evidence, and report-style synthesis.
Fable model traits
Four traits that change the research result
Each trait matters because it changes what appears in the final answer: fewer missed qualifiers, clearer tradeoffs, and stronger briefs.
01
ReadingContext-sensitive reading
Fable is suited to sources where the answer depends on context: who wrote it, what was measured, which definition is being used, and what the source leaves out.
Research effect
The report is less likely to flatten a nuanced source into a generic takeaway.
02
ReasoningMulti-hop evidence chaining
The model can carry a line of reasoning across several sources and use earlier evidence to test later claims.
Research effect
You get stronger answers for questions that cannot be solved by one search result.
03
JudgmentBuilt-in skepticism
Fable is valuable when a confident answer would be dangerous. It looks for weak support, missing evidence, and contradictions before smoothing the narrative.
Research effect
The final report can separate strong claims from claims that only sound strong.
04
SynthesisResearch memo composition
Instead of stopping at extracted notes, Fable organizes findings into a brief with conclusions, caveats, and follow-up questions.
Research effect
The output is easier to share with a team or use as decision material.
Better research outcomes
What improves when Fable leads the work
Use it when the cost of a shallow answer is higher than the cost of spending extra credits.
1
Answers survive a second read
Because qualifiers and source limits are visible, the result is easier to review and less likely to fall apart in discussion.
2
Contradictions become useful
Instead of hiding conflicting evidence, Fable can turn disagreement into a clearer map of what is known, uncertain, and worth checking next.
3
Follow-up work gets sharper
When the model exposes gaps, you can spend the next prompt on the exact missing piece instead of restarting the research.
How to brief Fable
Give the model a research assignment
A strong first prompt helps Fable spend credits on reasoning instead of guessing the shape of the task.
Brief move
State the decision
Tell Fable whether the output will support a product brief, investment memo, market scan, policy readout, or academic scope.
Brief move
Set evidence rules
Name source types that matter, sources to avoid, required date ranges, and whether confidence levels should be called out.
Brief move
Ask for weak points
Request contradictions, missing evidence, and assumptions before asking for the final report.
Credit reminder
Use the stronger model when the question deserves it
The Claude Fable model consumes credits quickly. Use it for research where deeper reading, conflict checks, and synthesis are worth the extra credits.
FAQ
Questions about Fable Research
Claude Fable is useful for research because it is better suited to careful source reading, multi-step reasoning, contradiction checks, and report synthesis. In Felo, those strengths are packaged as the Fable Research agent.